diff --git a/README.md b/README.md
index 4a5c5d299e8bff1afd3aa34c4b56cd0f32169e0e..b441e7bbeffb3ef6a181b2e829666fce0be8173a 100644
--- a/README.md
+++ b/README.md
@@ -76,7 +76,7 @@ Other features
- [x] [DCNv2](configs/dcn/README.md)
- [x] [Group Normalization](configs/gn/README.md)
- [x] [Weight Standardization](configs/gn+ws/README.md)
-- [x] [OHEM](configs/faster_rcnn_ohem_r50_fpn_1x.py)
+- [x] [OHEM](configs/faster_rcnn_ohem_r50_fpn_1x_coco.py)
- [x] Soft-NMS
- [x] [Generalized Attention](configs/empirical_attention/README.md)
- [x] [GCNet](configs/gcnet/README.md)
diff --git a/configs/_base_/datasets/cityscapes_detection.py b/configs/_base_/datasets/cityscapes_detection.py
new file mode 100644
index 0000000000000000000000000000000000000000..abbc082c46be7118206003855f0359a7492b974c
--- /dev/null
+++ b/configs/_base_/datasets/cityscapes_detection.py
@@ -0,0 +1,55 @@
+dataset_type = 'CityscapesDataset'
+data_root = 'data/cityscapes/'
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(
+ type='Resize', img_scale=[(2048, 800), (2048, 1024)], keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(2048, 1024),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ imgs_per_gpu=1,
+ workers_per_gpu=2,
+ train=dict(
+ type='RepeatDataset',
+ times=8,
+ dataset=dict(
+ type=dataset_type,
+ ann_file=data_root +
+ 'annotations/instancesonly_filtered_gtFine_train.json',
+ img_prefix=data_root + 'leftImg8bit/train/',
+ pipeline=train_pipeline)),
+ val=dict(
+ type=dataset_type,
+ ann_file=data_root +
+ 'annotations/instancesonly_filtered_gtFine_val.json',
+ img_prefix=data_root + 'leftImg8bit/val/',
+ pipeline=test_pipeline),
+ test=dict(
+ type=dataset_type,
+ ann_file=data_root +
+ 'annotations/instancesonly_filtered_gtFine_test.json',
+ img_prefix=data_root + 'leftImg8bit/test/',
+ pipeline=test_pipeline))
+evaluation = dict(interval=1, metric='bbox')
diff --git a/configs/_base_/datasets/cityscapes_instance.py b/configs/_base_/datasets/cityscapes_instance.py
new file mode 100644
index 0000000000000000000000000000000000000000..669132ebca25a524bff678df4623c42d8819bbb2
--- /dev/null
+++ b/configs/_base_/datasets/cityscapes_instance.py
@@ -0,0 +1,16 @@
+_base_ = './cityscapes_detection.py'
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
+ dict(
+ type='Resize', img_scale=[(2048, 800), (2048, 1024)], keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
+]
+data = dict(train=dict(dataset=dict(pipeline=train_pipeline)))
+evaluation = dict(metric=['bbox', 'segm'])
diff --git a/configs/_base_/datasets/coco_detection.py b/configs/_base_/datasets/coco_detection.py
new file mode 100644
index 0000000000000000000000000000000000000000..116a660a9c01c9c83857ea3a1ad258ef56800b99
--- /dev/null
+++ b/configs/_base_/datasets/coco_detection.py
@@ -0,0 +1,48 @@
+dataset_type = 'CocoDataset'
+data_root = 'data/coco/'
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(1333, 800),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ imgs_per_gpu=2,
+ workers_per_gpu=2,
+ train=dict(
+ type=dataset_type,
+ ann_file=data_root + 'annotations/instances_train2017.json',
+ img_prefix=data_root + 'train2017/',
+ pipeline=train_pipeline),
+ val=dict(
+ type=dataset_type,
+ ann_file=data_root + 'annotations/instances_val2017.json',
+ img_prefix=data_root + 'val2017/',
+ pipeline=test_pipeline),
+ test=dict(
+ type=dataset_type,
+ ann_file=data_root + 'annotations/instances_val2017.json',
+ img_prefix=data_root + 'val2017/',
+ pipeline=test_pipeline))
+evaluation = dict(interval=1, metric='bbox')
diff --git a/configs/_base_/datasets/coco_instance.py b/configs/_base_/datasets/coco_instance.py
new file mode 100644
index 0000000000000000000000000000000000000000..3d3f0102ce68b6ad0f1b81604f7397bb5310dda1
--- /dev/null
+++ b/configs/_base_/datasets/coco_instance.py
@@ -0,0 +1,15 @@
+_base_ = 'coco_detection.py'
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
+]
+data = dict(train=dict(pipeline=train_pipeline))
+evaluation = dict(metric=['bbox', 'segm'])
diff --git a/configs/_base_/datasets/coco_instance_semantic.py b/configs/_base_/datasets/coco_instance_semantic.py
new file mode 100644
index 0000000000000000000000000000000000000000..f4d124ba2956d7cc84f803fd8bf4ab53aa124d05
--- /dev/null
+++ b/configs/_base_/datasets/coco_instance_semantic.py
@@ -0,0 +1,40 @@
+_base_ = 'coco_detection.py'
+data_root = 'data/coco/'
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='LoadAnnotations', with_bbox=True, with_mask=True, with_seg=True),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='SegRescale', scale_factor=1 / 8),
+ dict(type='DefaultFormatBundle'),
+ dict(
+ type='Collect',
+ keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks', 'gt_semantic_seg']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(1333, 800),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ train=dict(
+ seg_prefix=data_root + 'stuffthingmaps/train2017/',
+ pipeline=train_pipeline),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+evaluation = dict(metric=['bbox', 'segm'])
diff --git a/configs/_base_/datasets/voc0712.py b/configs/_base_/datasets/voc0712.py
new file mode 100644
index 0000000000000000000000000000000000000000..fbb7ec820f096ec4f639332e5cc25ef669e3bb4a
--- /dev/null
+++ b/configs/_base_/datasets/voc0712.py
@@ -0,0 +1,55 @@
+# dataset settings
+dataset_type = 'VOCDataset'
+data_root = 'data/VOCdevkit/'
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(type='Resize', img_scale=(1000, 600), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(1000, 600),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ imgs_per_gpu=2,
+ workers_per_gpu=2,
+ train=dict(
+ type='RepeatDataset',
+ times=3,
+ dataset=dict(
+ type=dataset_type,
+ ann_file=[
+ data_root + 'VOC2007/ImageSets/Main/trainval.txt',
+ data_root + 'VOC2012/ImageSets/Main/trainval.txt'
+ ],
+ img_prefix=[data_root + 'VOC2007/', data_root + 'VOC2012/'],
+ pipeline=train_pipeline)),
+ val=dict(
+ type=dataset_type,
+ ann_file=data_root + 'VOC2007/ImageSets/Main/test.txt',
+ img_prefix=data_root + 'VOC2007/',
+ pipeline=test_pipeline),
+ test=dict(
+ type=dataset_type,
+ ann_file=data_root + 'VOC2007/ImageSets/Main/test.txt',
+ img_prefix=data_root + 'VOC2007/',
+ pipeline=test_pipeline))
+evaluation = dict(interval=1, metric='mAP')
diff --git a/configs/_base_/datasets/wider_face.py b/configs/_base_/datasets/wider_face.py
new file mode 100644
index 0000000000000000000000000000000000000000..c651c2cebebb2b142b35afe32875b04ac2efc226
--- /dev/null
+++ b/configs/_base_/datasets/wider_face.py
@@ -0,0 +1,63 @@
+# dataset settings
+dataset_type = 'WIDERFaceDataset'
+data_root = 'data/WIDERFace/'
+img_norm_cfg = dict(mean=[123.675, 116.28, 103.53], std=[1, 1, 1], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile', to_float32=True),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(
+ type='PhotoMetricDistortion',
+ brightness_delta=32,
+ contrast_range=(0.5, 1.5),
+ saturation_range=(0.5, 1.5),
+ hue_delta=18),
+ dict(
+ type='Expand',
+ mean=img_norm_cfg['mean'],
+ to_rgb=img_norm_cfg['to_rgb'],
+ ratio_range=(1, 4)),
+ dict(
+ type='MinIoURandomCrop',
+ min_ious=(0.1, 0.3, 0.5, 0.7, 0.9),
+ min_crop_size=0.3),
+ dict(type='Resize', img_scale=(300, 300), keep_ratio=False),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(300, 300),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=False),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ imgs_per_gpu=60,
+ workers_per_gpu=2,
+ train=dict(
+ type='RepeatDataset',
+ times=2,
+ dataset=dict(
+ type=dataset_type,
+ ann_file=data_root + 'train.txt',
+ img_prefix=data_root + 'WIDER_train/',
+ min_size=17,
+ pipeline=train_pipeline)),
+ val=dict(
+ type=dataset_type,
+ ann_file=data_root + 'val.txt',
+ img_prefix=data_root + 'WIDER_val/',
+ pipeline=test_pipeline),
+ test=dict(
+ type=dataset_type,
+ ann_file=data_root + 'val.txt',
+ img_prefix=data_root + 'WIDER_val/',
+ pipeline=test_pipeline))
diff --git a/configs/_base_/default_runtime.py b/configs/_base_/default_runtime.py
new file mode 100644
index 0000000000000000000000000000000000000000..594de8dcc99b9e4fc0208f327a05910a95a1793c
--- /dev/null
+++ b/configs/_base_/default_runtime.py
@@ -0,0 +1,14 @@
+checkpoint_config = dict(interval=1)
+# yapf:disable
+log_config = dict(
+ interval=50,
+ hooks=[
+ dict(type='TextLoggerHook'),
+ # dict(type='TensorboardLoggerHook')
+ ])
+# yapf:enable
+dist_params = dict(backend='nccl')
+log_level = 'INFO'
+load_from = None
+resume_from = None
+workflow = [('train', 1)]
diff --git a/configs/cascade_mask_rcnn_r50_fpn_1x.py b/configs/_base_/models/cascade_mask_rcnn_r50_fpn.py
similarity index 67%
rename from configs/cascade_mask_rcnn_r50_fpn_1x.py
rename to configs/_base_/models/cascade_mask_rcnn_r50_fpn.py
index 459eb17ae6d97804df623ad81aab777d8e06724c..3e86275a46cef8c59d5828645b0da99a02fb0ef3 100644
--- a/configs/cascade_mask_rcnn_r50_fpn_1x.py
+++ b/configs/_base_/models/cascade_mask_rcnn_r50_fpn.py
@@ -35,8 +35,7 @@ model = dict(
featmap_strides=[4, 8, 16, 32]),
bbox_head=[
dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
+ type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
@@ -48,8 +47,7 @@ model = dict(
type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
+ type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
@@ -61,8 +59,7 @@ model = dict(
type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
+ type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
@@ -176,79 +173,3 @@ test_cfg = dict(
nms=dict(type='nms', iou_thr=0.5),
max_per_img=100,
mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/cascade_mask_rcnn_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/cascade_rcnn_r50_fpn_1x.py b/configs/_base_/models/cascade_rcnn_r50_fpn.py
similarity index 65%
rename from configs/cascade_rcnn_r50_fpn_1x.py
rename to configs/_base_/models/cascade_rcnn_r50_fpn.py
index 6f7940307a70eee63822be18b85dc8a94e8dd4f6..274f8cbd7ce2d98dd3b33cc6716d90c40260ce63 100644
--- a/configs/cascade_rcnn_r50_fpn_1x.py
+++ b/configs/_base_/models/cascade_rcnn_r50_fpn.py
@@ -35,8 +35,7 @@ model = dict(
featmap_strides=[4, 8, 16, 32]),
bbox_head=[
dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
+ type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
@@ -48,8 +47,7 @@ model = dict(
type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
+ type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
@@ -61,8 +59,7 @@ model = dict(
type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
+ type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
@@ -157,79 +154,3 @@ test_cfg = dict(
min_bbox_size=0),
rcnn=dict(
score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/cascade_rcnn_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/_base_/models/fast_rcnn_r50_fpn.py b/configs/_base_/models/fast_rcnn_r50_fpn.py
new file mode 100644
index 0000000000000000000000000000000000000000..a6a7b17a0543ba6b52a349c9bb7f647af77c582e
--- /dev/null
+++ b/configs/_base_/models/fast_rcnn_r50_fpn.py
@@ -0,0 +1,54 @@
+# model settings
+model = dict(
+ type='FastRCNN',
+ pretrained='torchvision://resnet50',
+ backbone=dict(
+ type='ResNet',
+ depth=50,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'),
+ neck=dict(
+ type='FPN',
+ in_channels=[256, 512, 1024, 2048],
+ out_channels=256,
+ num_outs=5),
+ bbox_roi_extractor=dict(
+ type='SingleRoIExtractor',
+ roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
+ out_channels=256,
+ featmap_strides=[4, 8, 16, 32]),
+ bbox_head=dict(
+ type='Shared2FCBBoxHead',
+ in_channels=256,
+ fc_out_channels=1024,
+ roi_feat_size=7,
+ num_classes=81,
+ target_means=[0., 0., 0., 0.],
+ target_stds=[0.1, 0.1, 0.2, 0.2],
+ reg_class_agnostic=False,
+ loss_cls=dict(
+ type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
+ loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
+# model training and testing settings
+train_cfg = dict(
+ rcnn=dict(
+ assigner=dict(
+ type='MaxIoUAssigner',
+ pos_iou_thr=0.5,
+ neg_iou_thr=0.5,
+ min_pos_iou=0.5,
+ ignore_iof_thr=-1),
+ sampler=dict(
+ type='RandomSampler',
+ num=512,
+ pos_fraction=0.25,
+ neg_pos_ub=-1,
+ add_gt_as_proposals=True),
+ pos_weight=-1,
+ debug=False))
+test_cfg = dict(
+ rcnn=dict(
+ score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100))
diff --git a/configs/faster_rcnn_r50_caffe_c4_1x.py b/configs/_base_/models/faster_rcnn_r50_caffe_c4.py
similarity index 55%
rename from configs/faster_rcnn_r50_caffe_c4_1x.py
rename to configs/_base_/models/faster_rcnn_r50_caffe_c4.py
index fcac778c97eaca9a1d9dfd1c10a7a9bd10d32c53..14ca1ab1811c12f8b890a163675bd61fa281b2b4 100644
--- a/configs/faster_rcnn_r50_caffe_c4_1x.py
+++ b/configs/_base_/models/faster_rcnn_r50_caffe_c4.py
@@ -102,79 +102,3 @@ test_cfg = dict(
min_bbox_size=0),
rcnn=dict(
score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_r50_caffe_c4_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/faster_rcnn_r50_fpn_1x.py b/configs/_base_/models/faster_rcnn_r50_fpn.py
similarity index 54%
rename from configs/faster_rcnn_r50_fpn_1x.py
rename to configs/_base_/models/faster_rcnn_r50_fpn.py
index e1659fedb6138e2f9e7877a2948c52bc7ff24cff..05c6ad105c4cc4239a049b39ea4907df316a87a1 100644
--- a/configs/faster_rcnn_r50_fpn_1x.py
+++ b/configs/_base_/models/faster_rcnn_r50_fpn.py
@@ -1,4 +1,3 @@
-# model settings
model = dict(
type='FasterRCNN',
pretrained='torchvision://resnet50',
@@ -33,8 +32,7 @@ model = dict(
out_channels=256,
featmap_strides=[4, 8, 16, 32]),
bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
+ type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
@@ -98,79 +96,3 @@ test_cfg = dict(
# soft-nms is also supported for rcnn testing
# e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/mask_rcnn_r50_caffe_c4_1x.py b/configs/_base_/models/mask_rcnn_r50_caffe_c4.py
similarity index 57%
rename from configs/mask_rcnn_r50_caffe_c4_1x.py
rename to configs/_base_/models/mask_rcnn_r50_caffe_c4.py
index b5077715f2e6f7e3dfd1a08fd89b2a4068f82cb8..3f1c07bb8593c64da346e3c44ea4aef2fd04c36c 100644
--- a/configs/mask_rcnn_r50_caffe_c4_1x.py
+++ b/configs/_base_/models/mask_rcnn_r50_caffe_c4.py
@@ -2,7 +2,7 @@
norm_cfg = dict(type='BN', requires_grad=False)
model = dict(
type='MaskRCNN',
- # pretrained='open-mmlab://resnet50_caffe',
+ pretrained='open-mmlab://resnet50_caffe',
backbone=dict(
type='ResNet',
depth=50,
@@ -115,79 +115,3 @@ test_cfg = dict(
nms=dict(type='nms', iou_thr=0.5),
max_per_img=100,
mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_r50_caffe_c4_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/mask_rcnn_r50_fpn_1x.py b/configs/_base_/models/mask_rcnn_r50_fpn.py
similarity index 57%
rename from configs/mask_rcnn_r50_fpn_1x.py
rename to configs/_base_/models/mask_rcnn_r50_fpn.py
index b3a0e66a091f4c9a8f9f15504f15260c11cdb13f..dfa1a3465d0d432acb784b563089f2fb3d873630 100644
--- a/configs/mask_rcnn_r50_fpn_1x.py
+++ b/configs/_base_/models/mask_rcnn_r50_fpn.py
@@ -33,8 +33,7 @@ model = dict(
out_channels=256,
featmap_strides=[4, 8, 16, 32]),
bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
+ type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
@@ -112,79 +111,3 @@ test_cfg = dict(
nms=dict(type='nms', iou_thr=0.5),
max_per_img=100,
mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/_base_/models/retinanet_r50_fpn.py b/configs/_base_/models/retinanet_r50_fpn.py
new file mode 100644
index 0000000000000000000000000000000000000000..f94c27fe8ba6375b9af99fd445784dde626f7e73
--- /dev/null
+++ b/configs/_base_/models/retinanet_r50_fpn.py
@@ -0,0 +1,55 @@
+# model settings
+model = dict(
+ type='RetinaNet',
+ pretrained='torchvision://resnet50',
+ backbone=dict(
+ type='ResNet',
+ depth=50,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'),
+ neck=dict(
+ type='FPN',
+ in_channels=[256, 512, 1024, 2048],
+ out_channels=256,
+ start_level=1,
+ add_extra_convs=True,
+ num_outs=5),
+ bbox_head=dict(
+ type='RetinaHead',
+ num_classes=81,
+ in_channels=256,
+ stacked_convs=4,
+ feat_channels=256,
+ octave_base_scale=4,
+ scales_per_octave=3,
+ anchor_ratios=[0.5, 1.0, 2.0],
+ anchor_strides=[8, 16, 32, 64, 128],
+ target_means=[.0, .0, .0, .0],
+ target_stds=[1.0, 1.0, 1.0, 1.0],
+ loss_cls=dict(
+ type='FocalLoss',
+ use_sigmoid=True,
+ gamma=2.0,
+ alpha=0.25,
+ loss_weight=1.0),
+ loss_bbox=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0)))
+# training and testing settings
+train_cfg = dict(
+ assigner=dict(
+ type='MaxIoUAssigner',
+ pos_iou_thr=0.5,
+ neg_iou_thr=0.4,
+ min_pos_iou=0,
+ ignore_iof_thr=-1),
+ allowed_border=-1,
+ pos_weight=-1,
+ debug=False)
+test_cfg = dict(
+ nms_pre=1000,
+ min_bbox_size=0,
+ score_thr=0.05,
+ nms=dict(type='nms', iou_thr=0.5),
+ max_per_img=100)
diff --git a/configs/_base_/models/rpn_r50_caffe_c4.py b/configs/_base_/models/rpn_r50_caffe_c4.py
new file mode 100644
index 0000000000000000000000000000000000000000..ca9009e803f9f4d85ff6d576dbe35e4ced9cc204
--- /dev/null
+++ b/configs/_base_/models/rpn_r50_caffe_c4.py
@@ -0,0 +1,54 @@
+# model settings
+model = dict(
+ type='RPN',
+ pretrained='open-mmlab://resnet50_caffe',
+ backbone=dict(
+ type='ResNet',
+ depth=50,
+ num_stages=3,
+ strides=(1, 2, 2),
+ dilations=(1, 1, 1),
+ out_indices=(2, ),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=False),
+ norm_eval=True,
+ style='caffe'),
+ neck=None,
+ rpn_head=dict(
+ type='RPNHead',
+ in_channels=1024,
+ feat_channels=1024,
+ anchor_scales=[2, 4, 8, 16, 32],
+ anchor_ratios=[0.5, 1.0, 2.0],
+ anchor_strides=[16],
+ target_means=[.0, .0, .0, .0],
+ target_stds=[1.0, 1.0, 1.0, 1.0],
+ loss_cls=dict(
+ type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
+ loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)))
+# model training and testing settings
+train_cfg = dict(
+ rpn=dict(
+ assigner=dict(
+ type='MaxIoUAssigner',
+ pos_iou_thr=0.7,
+ neg_iou_thr=0.3,
+ min_pos_iou=0.3,
+ ignore_iof_thr=-1),
+ sampler=dict(
+ type='RandomSampler',
+ num=256,
+ pos_fraction=0.5,
+ neg_pos_ub=-1,
+ add_gt_as_proposals=False),
+ allowed_border=0,
+ pos_weight=-1,
+ debug=False))
+test_cfg = dict(
+ rpn=dict(
+ nms_across_levels=False,
+ nms_pre=12000,
+ nms_post=2000,
+ max_num=2000,
+ nms_thr=0.7,
+ min_bbox_size=0))
diff --git a/configs/_base_/models/rpn_r50_fpn.py b/configs/_base_/models/rpn_r50_fpn.py
new file mode 100644
index 0000000000000000000000000000000000000000..cd9503f9caebea19d41820f264a857c7c0c20b9e
--- /dev/null
+++ b/configs/_base_/models/rpn_r50_fpn.py
@@ -0,0 +1,55 @@
+# model settings
+model = dict(
+ type='RPN',
+ pretrained='torchvision://resnet50',
+ backbone=dict(
+ type='ResNet',
+ depth=50,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'),
+ neck=dict(
+ type='FPN',
+ in_channels=[256, 512, 1024, 2048],
+ out_channels=256,
+ num_outs=5),
+ rpn_head=dict(
+ type='RPNHead',
+ in_channels=256,
+ feat_channels=256,
+ anchor_scales=[8],
+ anchor_ratios=[0.5, 1.0, 2.0],
+ anchor_strides=[4, 8, 16, 32, 64],
+ target_means=[.0, .0, .0, .0],
+ target_stds=[1.0, 1.0, 1.0, 1.0],
+ loss_cls=dict(
+ type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
+ loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)))
+# model training and testing settings
+train_cfg = dict(
+ rpn=dict(
+ assigner=dict(
+ type='MaxIoUAssigner',
+ pos_iou_thr=0.7,
+ neg_iou_thr=0.3,
+ min_pos_iou=0.3,
+ ignore_iof_thr=-1),
+ sampler=dict(
+ type='RandomSampler',
+ num=256,
+ pos_fraction=0.5,
+ neg_pos_ub=-1,
+ add_gt_as_proposals=False),
+ allowed_border=0,
+ pos_weight=-1,
+ debug=False))
+test_cfg = dict(
+ rpn=dict(
+ nms_across_levels=False,
+ nms_pre=2000,
+ nms_post=2000,
+ max_num=2000,
+ nms_thr=0.7,
+ min_bbox_size=0))
diff --git a/configs/_base_/models/ssd300.py b/configs/_base_/models/ssd300.py
new file mode 100644
index 0000000000000000000000000000000000000000..66a8dc6ac3578c7ee3203e221f49f55e6f3c78d1
--- /dev/null
+++ b/configs/_base_/models/ssd300.py
@@ -0,0 +1,44 @@
+# model settings
+input_size = 300
+model = dict(
+ type='SingleStageDetector',
+ pretrained='open-mmlab://vgg16_caffe',
+ backbone=dict(
+ type='SSDVGG',
+ input_size=input_size,
+ depth=16,
+ with_last_pool=False,
+ ceil_mode=True,
+ out_indices=(3, 4),
+ out_feature_indices=(22, 34),
+ l2_norm_scale=20),
+ neck=None,
+ bbox_head=dict(
+ type='SSDHead',
+ input_size=input_size,
+ in_channels=(512, 1024, 512, 256, 256, 256),
+ num_classes=81,
+ anchor_strides=(8, 16, 32, 64, 100, 300),
+ basesize_ratio_range=(0.15, 0.9),
+ anchor_ratios=([2], [2, 3], [2, 3], [2, 3], [2], [2]),
+ target_means=(.0, .0, .0, .0),
+ target_stds=(0.1, 0.1, 0.2, 0.2)))
+cudnn_benchmark = True
+train_cfg = dict(
+ assigner=dict(
+ type='MaxIoUAssigner',
+ pos_iou_thr=0.5,
+ neg_iou_thr=0.5,
+ min_pos_iou=0.,
+ ignore_iof_thr=-1,
+ gt_max_assign_all=False),
+ smoothl1_beta=1.,
+ allowed_border=-1,
+ pos_weight=-1,
+ neg_pos_ratio=3,
+ debug=False)
+test_cfg = dict(
+ nms=dict(type='nms', iou_thr=0.45),
+ min_bbox_size=0,
+ score_thr=0.02,
+ max_per_img=200)
diff --git a/configs/_base_/schedules/schedule_1x.py b/configs/_base_/schedules/schedule_1x.py
new file mode 100644
index 0000000000000000000000000000000000000000..d559ec1d04f2f0e40d5e2c41e3ae763b7befde39
--- /dev/null
+++ b/configs/_base_/schedules/schedule_1x.py
@@ -0,0 +1,11 @@
+# optimizer
+optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
+optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
+# learning policy
+lr_config = dict(
+ policy='step',
+ warmup='linear',
+ warmup_iters=500,
+ warmup_ratio=1.0 / 3,
+ step=[8, 11])
+total_epochs = 12
diff --git a/configs/_base_/schedules/schedule_20e.py b/configs/_base_/schedules/schedule_20e.py
new file mode 100644
index 0000000000000000000000000000000000000000..e6ca2b24df28c4752c84dc76a2b2d45cdfbc200e
--- /dev/null
+++ b/configs/_base_/schedules/schedule_20e.py
@@ -0,0 +1,4 @@
+_base_ = './schedule_1x.py'
+# learning policy
+lr_config = dict(step=[16, 19])
+total_epochs = 20
diff --git a/configs/_base_/schedules/schedule_2x.py b/configs/_base_/schedules/schedule_2x.py
new file mode 100644
index 0000000000000000000000000000000000000000..72b4135c3d74a79dbad55b9ff24d2bddc88956db
--- /dev/null
+++ b/configs/_base_/schedules/schedule_2x.py
@@ -0,0 +1,4 @@
+_base_ = './schedule_1x.py'
+# learning policy
+lr_config = dict(step=[16, 22])
+total_epochs = 24
diff --git a/configs/albu_example/mask_rcnn_r50_fpn_1x.py b/configs/albu_example/mask_rcnn_r50_fpn_1x.py
deleted file mode 100644
index da3e23ae1ac6bdcac81863df91ab8d809e2ae321..0000000000000000000000000000000000000000
--- a/configs/albu_example/mask_rcnn_r50_fpn_1x.py
+++ /dev/null
@@ -1,249 +0,0 @@
-# model settings
-model = dict(
- type='MaskRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-albu_train_transforms = [
- dict(
- type='ShiftScaleRotate',
- shift_limit=0.0625,
- scale_limit=0.0,
- rotate_limit=0,
- interpolation=1,
- p=0.5),
- dict(
- type='RandomBrightnessContrast',
- brightness_limit=[0.1, 0.3],
- contrast_limit=[0.1, 0.3],
- p=0.2),
- dict(
- type='OneOf',
- transforms=[
- dict(
- type='RGBShift',
- r_shift_limit=10,
- g_shift_limit=10,
- b_shift_limit=10,
- p=1.0),
- dict(
- type='HueSaturationValue',
- hue_shift_limit=20,
- sat_shift_limit=30,
- val_shift_limit=20,
- p=1.0)
- ],
- p=0.1),
- dict(type='JpegCompression', quality_lower=85, quality_upper=95, p=0.2),
- dict(type='ChannelShuffle', p=0.1),
- dict(
- type='OneOf',
- transforms=[
- dict(type='Blur', blur_limit=3, p=1.0),
- dict(type='MedianBlur', blur_limit=3, p=1.0)
- ],
- p=0.1),
-]
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='Pad', size_divisor=32),
- dict(
- type='Albu',
- transforms=albu_train_transforms,
- bbox_params=dict(
- type='BboxParams',
- format='pascal_voc',
- label_fields=['gt_labels'],
- min_visibility=0.0,
- filter_lost_elements=True),
- keymap={
- 'img': 'image',
- 'gt_masks': 'masks',
- 'gt_bboxes': 'bboxes'
- },
- update_pad_shape=False,
- skip_img_without_anno=True),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='DefaultFormatBundle'),
- dict(
- type='Collect',
- keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks'],
- meta_keys=('filename', 'ori_shape', 'img_shape', 'img_norm_cfg',
- 'pad_shape', 'scale_factor'))
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/albu_example/mask_rcnn_r50_fpn_albu_1x_coco.py b/configs/albu_example/mask_rcnn_r50_fpn_albu_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..610dadd1574f8af30294b906b5fa34c9d1a77f5b
--- /dev/null
+++ b/configs/albu_example/mask_rcnn_r50_fpn_albu_1x_coco.py
@@ -0,0 +1,74 @@
+_base_ = '../mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py'
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+albu_train_transforms = [
+ dict(
+ type='ShiftScaleRotate',
+ shift_limit=0.0625,
+ scale_limit=0.0,
+ rotate_limit=0,
+ interpolation=1,
+ p=0.5),
+ dict(
+ type='RandomBrightnessContrast',
+ brightness_limit=[0.1, 0.3],
+ contrast_limit=[0.1, 0.3],
+ p=0.2),
+ dict(
+ type='OneOf',
+ transforms=[
+ dict(
+ type='RGBShift',
+ r_shift_limit=10,
+ g_shift_limit=10,
+ b_shift_limit=10,
+ p=1.0),
+ dict(
+ type='HueSaturationValue',
+ hue_shift_limit=20,
+ sat_shift_limit=30,
+ val_shift_limit=20,
+ p=1.0)
+ ],
+ p=0.1),
+ dict(type='JpegCompression', quality_lower=85, quality_upper=95, p=0.2),
+ dict(type='ChannelShuffle', p=0.1),
+ dict(
+ type='OneOf',
+ transforms=[
+ dict(type='Blur', blur_limit=3, p=1.0),
+ dict(type='MedianBlur', blur_limit=3, p=1.0)
+ ],
+ p=0.1),
+]
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='Pad', size_divisor=32),
+ dict(
+ type='Albu',
+ transforms=albu_train_transforms,
+ bbox_params=dict(
+ type='BboxParams',
+ format='pascal_voc',
+ label_fields=['gt_labels'],
+ min_visibility=0.0,
+ filter_lost_elements=True),
+ keymap={
+ 'img': 'image',
+ 'gt_masks': 'masks',
+ 'gt_bboxes': 'bboxes'
+ },
+ update_pad_shape=False,
+ skip_img_without_anno=True),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='DefaultFormatBundle'),
+ dict(
+ type='Collect',
+ keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks'],
+ meta_keys=('filename', 'ori_shape', 'img_shape', 'img_norm_cfg',
+ 'pad_shape', 'scale_factor'))
+]
+data = dict(train=dict(pipeline=train_pipeline))
+work_dir = './work_dirs/mask_rcnn_r50_fpn_1x'
diff --git a/configs/atss/atss_r50_fpn.py b/configs/atss/atss_r50_fpn.py
new file mode 100644
index 0000000000000000000000000000000000000000..67b514789ca00c52bdb9251446058955de7739a2
--- /dev/null
+++ b/configs/atss/atss_r50_fpn.py
@@ -0,0 +1,52 @@
+model = dict(
+ type='ATSS',
+ pretrained='torchvision://resnet50',
+ backbone=dict(
+ type='ResNet',
+ depth=50,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'),
+ neck=dict(
+ type='FPN',
+ in_channels=[256, 512, 1024, 2048],
+ out_channels=256,
+ start_level=1,
+ add_extra_convs=True,
+ extra_convs_on_inputs=False,
+ num_outs=5),
+ bbox_head=dict(
+ type='ATSSHead',
+ num_classes=81,
+ in_channels=256,
+ stacked_convs=4,
+ feat_channels=256,
+ octave_base_scale=8,
+ scales_per_octave=1,
+ anchor_ratios=[1.0],
+ anchor_strides=[8, 16, 32, 64, 128],
+ target_means=[.0, .0, .0, .0],
+ target_stds=[0.1, 0.1, 0.2, 0.2],
+ loss_cls=dict(
+ type='FocalLoss',
+ use_sigmoid=True,
+ gamma=2.0,
+ alpha=0.25,
+ loss_weight=1.0),
+ loss_bbox=dict(type='GIoULoss', loss_weight=2.0),
+ loss_centerness=dict(
+ type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0)))
+# training and testing settings
+train_cfg = dict(
+ assigner=dict(type='ATSSAssigner', topk=9),
+ allowed_border=-1,
+ pos_weight=-1,
+ debug=False)
+test_cfg = dict(
+ nms_pre=1000,
+ min_bbox_size=0,
+ score_thr=0.05,
+ nms=dict(type='nms', iou_thr=0.6),
+ max_per_img=100)
diff --git a/configs/atss/atss_r50_fpn_1x.py b/configs/atss/atss_r50_fpn_1x.py
deleted file mode 100644
index f65e7ef386b28f1f44085d29be734892374d23d6..0000000000000000000000000000000000000000
--- a/configs/atss/atss_r50_fpn_1x.py
+++ /dev/null
@@ -1,129 +0,0 @@
-# model settings
-model = dict(
- type='ATSS',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- extra_convs_on_inputs=False,
- num_outs=5),
- bbox_head=dict(
- type='ATSSHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- octave_base_scale=8,
- scales_per_octave=1,
- anchor_ratios=[1.0],
- anchor_strides=[8, 16, 32, 64, 128],
- target_means=[.0, .0, .0, .0],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox=dict(type='GIoULoss', loss_weight=2.0),
- loss_centerness=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict(
- assigner=dict(type='ATSSAssigner', topk=9),
- allowed_border=-1,
- pos_weight=-1,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.6),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/atss_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/atss/atss_r50_fpn_1x_coco.py b/configs/atss/atss_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..4e7ae0d07126ada20d80b20893100c263d21c7bc
--- /dev/null
+++ b/configs/atss/atss_r50_fpn_1x_coco.py
@@ -0,0 +1,7 @@
+_base_ = [
+ 'atss_r50_fpn.py', '../_base_/datasets/coco_detection.py',
+ '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
+]
+# optimizer
+optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
+work_dir = './work_dirs/atss_r50_fpn_1x'
diff --git a/configs/carafe/faster_rcnn_r50_fpn_carafe_1x.py b/configs/carafe/faster_rcnn_r50_fpn_carafe_1x.py
deleted file mode 100644
index 111bc4e6ef39e430c236b2180e270b3d1ba9cded..0000000000000000000000000000000000000000
--- a/configs/carafe/faster_rcnn_r50_fpn_carafe_1x.py
+++ /dev/null
@@ -1,188 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN_CARAFE',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5,
- start_level=0,
- end_level=-1,
- norm_cfg=None,
- act_cfg=None,
- order=('conv', 'norm', 'act'),
- upsample_cfg=dict(
- type='carafe',
- up_kernel=5,
- up_group=1,
- encoder_kernel=3,
- encoder_dilation=1,
- compressed_channels=64)),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=64),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=64),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_r50_fpn_carafe_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/carafe/faster_rcnn_r50_fpn_carafe_1x_coco.py b/configs/carafe/faster_rcnn_r50_fpn_carafe_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..77814b5a2219e20ec50e3a48190bb4cb8f93fe24
--- /dev/null
+++ b/configs/carafe/faster_rcnn_r50_fpn_carafe_1x_coco.py
@@ -0,0 +1,51 @@
+_base_ = '../faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ neck=dict(
+ type='FPN_CARAFE',
+ in_channels=[256, 512, 1024, 2048],
+ out_channels=256,
+ num_outs=5,
+ start_level=0,
+ end_level=-1,
+ norm_cfg=None,
+ act_cfg=None,
+ order=('conv', 'norm', 'act'),
+ upsample_cfg=dict(
+ type='carafe',
+ up_kernel=5,
+ up_group=1,
+ encoder_kernel=3,
+ encoder_dilation=1,
+ compressed_channels=64)))
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=64),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(1333, 800),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=64),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ train=dict(pipeline=train_pipeline),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+work_dir = './work_dirs/faster_rcnn_r50_fpn_carafe_1x'
diff --git a/configs/carafe/mask_rcnn_r50_fpn_carafe_1x.py b/configs/carafe/mask_rcnn_r50_fpn_carafe_1x.py
deleted file mode 100644
index 67e3319514c3a4db0169c7bec313c116375a1cda..0000000000000000000000000000000000000000
--- a/configs/carafe/mask_rcnn_r50_fpn_carafe_1x.py
+++ /dev/null
@@ -1,210 +0,0 @@
-# model settings
-model = dict(
- type='MaskRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN_CARAFE',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5,
- start_level=0,
- end_level=-1,
- norm_cfg=None,
- act_cfg=None,
- order=('conv', 'norm', 'act'),
- upsample_cfg=dict(
- type='carafe',
- up_kernel=5,
- up_group=1,
- encoder_kernel=3,
- encoder_dilation=1,
- compressed_channels=64)),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- upsample_cfg=dict(
- type='carafe',
- scale_factor=2,
- up_kernel=5,
- up_group=1,
- encoder_kernel=3,
- encoder_dilation=1,
- compressed_channels=64),
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=64),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=64),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_r50_fpn_carafe_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/carafe/mask_rcnn_r50_fpn_carafe_1x_coco.py b/configs/carafe/mask_rcnn_r50_fpn_carafe_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..27ac8e77997f485f43ccea314e51140728f028ef
--- /dev/null
+++ b/configs/carafe/mask_rcnn_r50_fpn_carafe_1x_coco.py
@@ -0,0 +1,60 @@
+_base_ = '../mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ neck=dict(
+ type='FPN_CARAFE',
+ in_channels=[256, 512, 1024, 2048],
+ out_channels=256,
+ num_outs=5,
+ start_level=0,
+ end_level=-1,
+ norm_cfg=None,
+ act_cfg=None,
+ order=('conv', 'norm', 'act'),
+ upsample_cfg=dict(
+ type='carafe',
+ up_kernel=5,
+ up_group=1,
+ encoder_kernel=3,
+ encoder_dilation=1,
+ compressed_channels=64)),
+ mask_head=dict(
+ upsample_cfg=dict(
+ type='carafe',
+ scale_factor=2,
+ up_kernel=5,
+ up_group=1,
+ encoder_kernel=3,
+ encoder_dilation=1,
+ compressed_channels=64)))
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=64),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(1333, 800),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=64),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ train=dict(pipeline=train_pipeline),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+work_dir = './work_dirs/mask_rcnn_r50_fpn_carafe_1x'
diff --git a/configs/cascade_mask_rcnn_r101_fpn_1x.py b/configs/cascade_mask_rcnn_r101_fpn_1x.py
deleted file mode 100644
index d5de4a56e7721ed013c7a1c0285e044a5ed136c7..0000000000000000000000000000000000000000
--- a/configs/cascade_mask_rcnn_r101_fpn_1x.py
+++ /dev/null
@@ -1,254 +0,0 @@
-# model settings
-model = dict(
- type='CascadeRCNN',
- num_stages=3,
- pretrained='torchvision://resnet101',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=[
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
- ],
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/cascade_mask_rcnn_r101_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/cascade_mask_rcnn_r50_caffe_c4_1x.py b/configs/cascade_mask_rcnn_r50_caffe_c4_1x.py
deleted file mode 100644
index 867ef7cb56aab017202c2facf676b8686957795d..0000000000000000000000000000000000000000
--- a/configs/cascade_mask_rcnn_r50_caffe_c4_1x.py
+++ /dev/null
@@ -1,255 +0,0 @@
-# model settings
-norm_cfg = dict(type='BN', requires_grad=False)
-model = dict(
- type='CascadeRCNN',
- num_stages=3,
- pretrained='open-mmlab://resnet50_caffe',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=3,
- strides=(1, 2, 2),
- dilations=(1, 1, 1),
- out_indices=(2, ),
- frozen_stages=1,
- norm_cfg=norm_cfg,
- norm_eval=True,
- style='caffe'),
- shared_head=dict(
- type='ResLayer',
- depth=50,
- stage=3,
- stride=2,
- dilation=1,
- style='caffe',
- norm_cfg=norm_cfg,
- norm_eval=True),
- rpn_head=dict(
- type='RPNHead',
- in_channels=1024,
- feat_channels=1024,
- anchor_scales=[2, 4, 8, 16, 32],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[16],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=1024,
- featmap_strides=[16]),
- bbox_head=[
- dict(
- type='BBoxHead',
- with_avg_pool=True,
- roi_feat_size=7,
- in_channels=2048,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='BBoxHead',
- with_avg_pool=True,
- roi_feat_size=7,
- in_channels=2048,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='BBoxHead',
- with_avg_pool=True,
- roi_feat_size=7,
- in_channels=2048,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
- ],
- mask_roi_extractor=None,
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=0,
- in_channels=2048,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=12000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=14,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=14,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=14,
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=6000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/cascade_mask_rcnn_r50_caffe_c4_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/cascade_mask_rcnn_x101_32x4d_fpn_1x.py b/configs/cascade_mask_rcnn_x101_32x4d_fpn_1x.py
deleted file mode 100644
index 4a1edbfd0a3fee44c30a321122252b23d97c0299..0000000000000000000000000000000000000000
--- a/configs/cascade_mask_rcnn_x101_32x4d_fpn_1x.py
+++ /dev/null
@@ -1,256 +0,0 @@
-# model settings
-model = dict(
- type='CascadeRCNN',
- num_stages=3,
- pretrained='open-mmlab://resnext101_32x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=32,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=[
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
- ],
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/cascade_mask_rcnn_x101_32x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/cascade_mask_rcnn_x101_64x4d_fpn_1x.py b/configs/cascade_mask_rcnn_x101_64x4d_fpn_1x.py
deleted file mode 100644
index d4d1f09bb90396c36e21ef4f8ac47a3921f42396..0000000000000000000000000000000000000000
--- a/configs/cascade_mask_rcnn_x101_64x4d_fpn_1x.py
+++ /dev/null
@@ -1,256 +0,0 @@
-# model settings
-model = dict(
- type='CascadeRCNN',
- num_stages=3,
- pretrained='open-mmlab://resnext101_64x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=64,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=[
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
- ],
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/cascade_mask_rcnn_x101_64x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/cascade_rcnn/cascade_mask_rcnn_r101_fpn_1x_coco.py b/configs/cascade_rcnn/cascade_mask_rcnn_r101_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..888b70b3f08062dfce9809f9e1f24268fa6ed290
--- /dev/null
+++ b/configs/cascade_rcnn/cascade_mask_rcnn_r101_fpn_1x_coco.py
@@ -0,0 +1,3 @@
+_base_ = './cascade_mask_rcnn_r50_fpn_1x_coco.py'
+model = dict(pretrained='torchvision://resnet101', backbone=dict(depth=101))
+work_dir = './work_dirs/cascade_mask_rcnn_r101_fpn_1x'
diff --git a/configs/cascade_rcnn/cascade_mask_rcnn_r50_fpn_1x_coco.py b/configs/cascade_rcnn/cascade_mask_rcnn_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..dd17860f1fc7545ead5ba1cd7a3712578dc9f108
--- /dev/null
+++ b/configs/cascade_rcnn/cascade_mask_rcnn_r50_fpn_1x_coco.py
@@ -0,0 +1,6 @@
+_base_ = [
+ '../_base_/models/cascade_mask_rcnn_r50_fpn.py',
+ '../_base_/datasets/coco_instance.py',
+ '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
+]
+work_dir = './work_dirs/cascade_mask_rcnn_r50_fpn_1x'
diff --git a/configs/cascade_rcnn/cascade_mask_rcnn_x101_32x4d_fpn_1x_coco.py b/configs/cascade_rcnn/cascade_mask_rcnn_x101_32x4d_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..0b0df7594050a4560e8b94f77804d6044981fc1d
--- /dev/null
+++ b/configs/cascade_rcnn/cascade_mask_rcnn_x101_32x4d_fpn_1x_coco.py
@@ -0,0 +1,14 @@
+_base_ = './cascade_mask_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_32x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=32,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+work_dir = './work_dirs/cascade_mask_rcnn_x101_32x4d_fpn_1x'
diff --git a/configs/cascade_rcnn/cascade_mask_rcnn_x101_64x4d_fpn_1x_coco.py b/configs/cascade_rcnn/cascade_mask_rcnn_x101_64x4d_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..7b4829bda32fa0e2dc48b062f5ab61ebe4bc5dbc
--- /dev/null
+++ b/configs/cascade_rcnn/cascade_mask_rcnn_x101_64x4d_fpn_1x_coco.py
@@ -0,0 +1,14 @@
+_base_ = './cascade_mask_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_64x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=64,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+work_dir = './work_dirs/cascade_mask_rcnn_x101_64x4d_fpn_1x'
diff --git a/configs/cascade_rcnn/cascade_rcnn_r101_fpn_1x_coco.py b/configs/cascade_rcnn/cascade_rcnn_r101_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..65093207fb0531da29cae6f357bd4cd9d2e794be
--- /dev/null
+++ b/configs/cascade_rcnn/cascade_rcnn_r101_fpn_1x_coco.py
@@ -0,0 +1,3 @@
+_base_ = './cascade_rcnn_r50_fpn_1x_coco.py'
+model = dict(pretrained='torchvision://resnet101', backbone=dict(depth=101))
+work_dir = './work_dirs/cascade_rcnn_r101_fpn_1x'
diff --git a/configs/cascade_rcnn/cascade_rcnn_r50_fpn_1x_coco.py b/configs/cascade_rcnn/cascade_rcnn_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..e6f01cb4364ef77a340510141d00d5a8b19140bb
--- /dev/null
+++ b/configs/cascade_rcnn/cascade_rcnn_r50_fpn_1x_coco.py
@@ -0,0 +1,6 @@
+_base_ = [
+ '../_base_/models/cascade_rcnn_r50_fpn.py',
+ '../_base_/datasets/coco_detection.py',
+ '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
+]
+work_dir = './work_dirs/cascade_rcnn_r50_fpn_1x'
diff --git a/configs/cascade_rcnn/cascade_rcnn_x101_32x4d_fpn_1x_coco.py b/configs/cascade_rcnn/cascade_rcnn_x101_32x4d_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..c39afb299112eea7d22103706c77e081268c942d
--- /dev/null
+++ b/configs/cascade_rcnn/cascade_rcnn_x101_32x4d_fpn_1x_coco.py
@@ -0,0 +1,14 @@
+_base_ = './cascade_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_32x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=32,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+work_dir = './work_dirs/cascade_rcnn_x101_32x4d_fpn_1x'
diff --git a/configs/cascade_rcnn/cascade_rcnn_x101_64x4d_fpn_1x_coco.py b/configs/cascade_rcnn/cascade_rcnn_x101_64x4d_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..496509496be524a9ddfadf8f5435dde2396a54eb
--- /dev/null
+++ b/configs/cascade_rcnn/cascade_rcnn_x101_64x4d_fpn_1x_coco.py
@@ -0,0 +1,16 @@
+_base_ = './cascade_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ type='CascadeRCNN',
+ num_stages=3,
+ pretrained='open-mmlab://resnext101_64x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=64,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+work_dir = './work_dirs/cascade_rcnn_x101_64x4d_fpn_1x'
diff --git a/configs/cascade_rcnn_r101_fpn_1x.py b/configs/cascade_rcnn_r101_fpn_1x.py
deleted file mode 100644
index 29fd1b7076eb8a8fecd5d12a7e4be107aa214cfe..0000000000000000000000000000000000000000
--- a/configs/cascade_rcnn_r101_fpn_1x.py
+++ /dev/null
@@ -1,235 +0,0 @@
-# model settings
-model = dict(
- type='CascadeRCNN',
- num_stages=3,
- pretrained='torchvision://resnet101',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=[
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
- ])
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/cascade_rcnn_r101_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/cascade_rcnn_r50_caffe_c4_1x.py b/configs/cascade_rcnn_r50_caffe_c4_1x.py
deleted file mode 100644
index 926310e27f7e14e40206b88e762d3c56864e4fbb..0000000000000000000000000000000000000000
--- a/configs/cascade_rcnn_r50_caffe_c4_1x.py
+++ /dev/null
@@ -1,243 +0,0 @@
-# model settings
-norm_cfg = dict(type='BN', requires_grad=False)
-model = dict(
- type='CascadeRCNN',
- num_stages=3,
- pretrained='open-mmlab://resnet50_caffe',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=3,
- strides=(1, 2, 2),
- dilations=(1, 1, 1),
- out_indices=(2, ),
- frozen_stages=1,
- norm_cfg=norm_cfg,
- norm_eval=True,
- style='caffe'),
- shared_head=dict(
- type='ResLayer',
- depth=50,
- stage=3,
- stride=2,
- dilation=1,
- style='caffe',
- norm_cfg=norm_cfg,
- norm_eval=True),
- rpn_head=dict(
- type='RPNHead',
- in_channels=1024,
- feat_channels=1024,
- anchor_scales=[2, 4, 8, 16, 32],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[16],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=1024,
- featmap_strides=[16]),
- bbox_head=[
- dict(
- type='BBoxHead',
- with_avg_pool=True,
- roi_feat_size=7,
- in_channels=2048,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='BBoxHead',
- with_avg_pool=True,
- roi_feat_size=7,
- in_channels=2048,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='BBoxHead',
- with_avg_pool=True,
- roi_feat_size=7,
- in_channels=2048,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- ])
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=12000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=14,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=14,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=14,
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=6000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/cascade_rcnn_r50_c4_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/cascade_rcnn_x101_32x4d_fpn_1x.py b/configs/cascade_rcnn_x101_32x4d_fpn_1x.py
deleted file mode 100644
index 4806dea90b7a3e1bbeeafc52f1b7529d82b4bee1..0000000000000000000000000000000000000000
--- a/configs/cascade_rcnn_x101_32x4d_fpn_1x.py
+++ /dev/null
@@ -1,237 +0,0 @@
-# model settings
-model = dict(
- type='CascadeRCNN',
- num_stages=3,
- pretrained='open-mmlab://resnext101_32x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=32,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=[
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
- ])
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/cascade_rcnn_x101_32x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/cascade_rcnn_x101_64x4d_fpn_1x.py b/configs/cascade_rcnn_x101_64x4d_fpn_1x.py
deleted file mode 100644
index 1d2667a85bc773122b3a3464e87f0a1b4259a4ea..0000000000000000000000000000000000000000
--- a/configs/cascade_rcnn_x101_64x4d_fpn_1x.py
+++ /dev/null
@@ -1,237 +0,0 @@
-# model settings
-model = dict(
- type='CascadeRCNN',
- num_stages=3,
- pretrained='open-mmlab://resnext101_64x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=64,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=[
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
- ])
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/cascade_rcnn_x101_64x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/cityscapes/faster_rcnn_r50_fpn_1x_cityscapes.py b/configs/cityscapes/faster_rcnn_r50_fpn_1x_cityscapes.py
index 4fd17e302187f5d6524a16354c6114a4b84ac469..82a5ec2a906e9738812eb1af9afdbda332807173 100644
--- a/configs/cityscapes/faster_rcnn_r50_fpn_1x_cityscapes.py
+++ b/configs/cityscapes/faster_rcnn_r50_fpn_1x_cityscapes.py
@@ -1,40 +1,12 @@
-# model settings
+_base_ = [
+ '../_base_/models/faster_rcnn_r50_fpn.py',
+ '../_base_/datasets/cityscapes_detection.py',
+ '../_base_/default_runtime.py'
+]
model = dict(
- type='FasterRCNN',
pretrained=None,
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
+ type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
@@ -45,115 +17,6 @@ model = dict(
loss_cls=dict(
type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CityscapesDataset'
-data_root = 'data/cityscapes/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(
- type='Resize', img_scale=[(2048, 800), (2048, 1024)], keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(2048, 1024),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=1,
- workers_per_gpu=2,
- train=dict(
- type='RepeatDataset',
- times=8,
- dataset=dict(
- type=dataset_type,
- ann_file=data_root +
- 'annotations/instancesonly_filtered_gtFine_train.json',
- img_prefix=data_root + 'leftImg8bit/train/',
- pipeline=train_pipeline)),
- val=dict(
- type=dataset_type,
- ann_file=data_root +
- 'annotations/instancesonly_filtered_gtFine_val.json',
- img_prefix=data_root + 'leftImg8bit/val/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root +
- 'annotations/instancesonly_filtered_gtFine_test.json',
- img_prefix=data_root + 'leftImg8bit/test/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
# optimizer
# lr is set for a batch size of 8
optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
@@ -166,21 +29,8 @@ lr_config = dict(
warmup_ratio=1.0 / 3,
# [7] yields higher performance than [6]
step=[7])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=100,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
total_epochs = 8 # actual epoch = 8 * 8 = 64
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
+log_config = dict(interval=100)
work_dir = './work_dirs/cityscapes/faster_rcnn_r50_fpn_1x_cityscapes'
# For better, more stable performance initialize from COCO
load_from = 'https://s3.ap-northeast-2.amazonaws.com/open-mmlab/mmdetection/models/faster_rcnn_r50_fpn_2x_20181010-443129e1.pth' # noqa
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/cityscapes/mask_rcnn_r50_fpn_1x_cityscapes.py b/configs/cityscapes/mask_rcnn_r50_fpn_1x_cityscapes.py
index e6481c7aa8ec5b1e6db0563310a628e8a3ccc731..39501a43e55cc81c55c7b5b59a6d46e7c05c0ad2 100644
--- a/configs/cityscapes/mask_rcnn_r50_fpn_1x_cityscapes.py
+++ b/configs/cityscapes/mask_rcnn_r50_fpn_1x_cityscapes.py
@@ -1,40 +1,11 @@
-# model settings
+_base_ = [
+ '../_base_/models/mask_rcnn_r50_fpn.py',
+ '../_base_/datasets/cityscapes_instance.py', '../_base_/default_runtime.py'
+]
model = dict(
- type='MaskRCNN',
pretrained=None,
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
+ type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
@@ -45,11 +16,6 @@ model = dict(
loss_cls=dict(
type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
mask_head=dict(
type='FCNMaskHead',
num_convs=4,
@@ -58,116 +24,6 @@ model = dict(
num_classes=9,
loss_mask=dict(
type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CityscapesDataset'
-data_root = 'data/cityscapes/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(
- type='Resize', img_scale=[(2048, 800), (2048, 1024)], keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(2048, 1024),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=1,
- workers_per_gpu=2,
- train=dict(
- type='RepeatDataset',
- times=8,
- dataset=dict(
- type=dataset_type,
- ann_file=data_root +
- 'annotations/instancesonly_filtered_gtFine_train.json',
- img_prefix=data_root + 'leftImg8bit/train/',
- pipeline=train_pipeline)),
- val=dict(
- type=dataset_type,
- ann_file=data_root +
- 'annotations/instancesonly_filtered_gtFine_val.json',
- img_prefix=data_root + 'leftImg8bit/val/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root +
- 'annotations/instancesonly_filtered_gtFine_test.json',
- img_prefix=data_root + 'leftImg8bit/test/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
# optimizer
# lr is set for a batch size of 8
optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
@@ -180,21 +36,8 @@ lr_config = dict(
warmup_ratio=1.0 / 3,
# [7] yields higher performance than [6]
step=[7])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=100,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
total_epochs = 8 # actual epoch = 8 * 8 = 64
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
+log_config = dict(interval=100)
work_dir = './work_dirs/cityscapes/mask_rcnn_r50_fpn_1x_cityscapes'
# For better, more stable performance initialize from COCO
load_from = 'https://s3.ap-northeast-2.amazonaws.com/open-mmlab/mmdetection/models/mask_rcnn_r50_fpn_2x_20181010-41d35c05.pth' # noqa
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/dcn/cascade_mask_rcnn_dconv_c3-c5_r50_fpn_1x.py b/configs/dcn/cascade_mask_rcnn_dconv_c3-c5_r50_fpn_1x.py
deleted file mode 100644
index 7d82c6b2b989f4599fbe1b59793052dade1a8e1d..0000000000000000000000000000000000000000
--- a/configs/dcn/cascade_mask_rcnn_dconv_c3-c5_r50_fpn_1x.py
+++ /dev/null
@@ -1,256 +0,0 @@
-# model settings
-model = dict(
- type='CascadeRCNN',
- num_stages=3,
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
- stage_with_dcn=(False, True, True, True)),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=[
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
- ],
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/cascade_mask_rcnn_dconv_c3-c5_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/dcn/cascade_mask_rcnn_r50_fpn_dconv_c3-c5_1x_coco.py b/configs/dcn/cascade_mask_rcnn_r50_fpn_dconv_c3-c5_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..04346844f2cf0d788a575b36979a3ed83b9d0333
--- /dev/null
+++ b/configs/dcn/cascade_mask_rcnn_r50_fpn_dconv_c3-c5_1x_coco.py
@@ -0,0 +1,6 @@
+_base_ = '../cascade_rcnn/cascade_mask_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ backbone=dict(
+ dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
+ stage_with_dcn=(False, True, True, True)))
+work_dir = './work_dirs/cascade_mask_rcnn_dconv_c3-c5_r50_fpn_1x'
diff --git a/configs/dcn/cascade_rcnn_dconv_c3-c5_r50_fpn_1x.py b/configs/dcn/cascade_rcnn_dconv_c3-c5_r50_fpn_1x.py
deleted file mode 100644
index 9ac1f985fa1e64407ae0883d9752d1ca3fc2f3bb..0000000000000000000000000000000000000000
--- a/configs/dcn/cascade_rcnn_dconv_c3-c5_r50_fpn_1x.py
+++ /dev/null
@@ -1,237 +0,0 @@
-# model settings
-model = dict(
- type='CascadeRCNN',
- num_stages=3,
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
- stage_with_dcn=(False, True, True, True)),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=[
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
- ])
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/cascade_rcnn_dconv_c3-c5_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/dcn/cascade_rcnn_r50_fpn_dconv_c3-c5_1x_coco.py b/configs/dcn/cascade_rcnn_r50_fpn_dconv_c3-c5_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..6337f08efb70f2a98d23de140e072c03b1da781b
--- /dev/null
+++ b/configs/dcn/cascade_rcnn_r50_fpn_dconv_c3-c5_1x_coco.py
@@ -0,0 +1,6 @@
+_base_ = '../cascade_rcnn/cascade_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ backbone=dict(
+ dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
+ stage_with_dcn=(False, True, True, True)))
+work_dir = './work_dirs/cascade_rcnn_dconv_c3-c5_r50_fpn_1x'
diff --git a/configs/dcn/faster_rcnn_dconv_c3-c5_r50_fpn_1x.py b/configs/dcn/faster_rcnn_dconv_c3-c5_r50_fpn_1x.py
deleted file mode 100644
index b252251f435cffe7de031af2d4ebe2f614461fee..0000000000000000000000000000000000000000
--- a/configs/dcn/faster_rcnn_dconv_c3-c5_r50_fpn_1x.py
+++ /dev/null
@@ -1,178 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
- stage_with_dcn=(False, True, True, True)),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_dconv_c3-c5_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/dcn/faster_rcnn_dconv_c3-c5_x101_32x4d_fpn_1x.py b/configs/dcn/faster_rcnn_dconv_c3-c5_x101_32x4d_fpn_1x.py
deleted file mode 100644
index 53a07c3345ce9383559c92736ea7d7b776abcb97..0000000000000000000000000000000000000000
--- a/configs/dcn/faster_rcnn_dconv_c3-c5_x101_32x4d_fpn_1x.py
+++ /dev/null
@@ -1,180 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='open-mmlab://resnext101_32x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=32,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
- stage_with_dcn=(False, True, True, True)),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_dconv_c3-c5_x101_32x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/dcn/faster_rcnn_dpool_r50_fpn_1x.py b/configs/dcn/faster_rcnn_dpool_r50_fpn_1x.py
deleted file mode 100644
index 124ba04653f2b0c5abf6dc6c73c7ad55a6b488f0..0000000000000000000000000000000000000000
--- a/configs/dcn/faster_rcnn_dpool_r50_fpn_1x.py
+++ /dev/null
@@ -1,182 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(
- type='DeformRoIPoolingPack',
- out_size=7,
- out_channels=256,
- no_trans=False,
- group_size=1,
- trans_std=0.1),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_dpool_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/dcn/faster_rcnn_mdconv_c3-c5_group4_r50_fpn_1x.py b/configs/dcn/faster_rcnn_mdconv_c3-c5_group4_r50_fpn_1x.py
deleted file mode 100644
index e1d2cddcb142ab3f5792041e233a951b1352da90..0000000000000000000000000000000000000000
--- a/configs/dcn/faster_rcnn_mdconv_c3-c5_group4_r50_fpn_1x.py
+++ /dev/null
@@ -1,178 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- dcn=dict(type='DCNv2', deformable_groups=4, fallback_on_stride=False),
- stage_with_dcn=(False, True, True, True)),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_mdconv_c3-c5_group4_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/dcn/faster_rcnn_mdconv_c3-c5_r50_fpn_1x.py b/configs/dcn/faster_rcnn_mdconv_c3-c5_r50_fpn_1x.py
deleted file mode 100644
index e22dbd4391212f3113cc1e98b25af66198411f35..0000000000000000000000000000000000000000
--- a/configs/dcn/faster_rcnn_mdconv_c3-c5_r50_fpn_1x.py
+++ /dev/null
@@ -1,178 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- dcn=dict(type='DCNv2', deformable_groups=1, fallback_on_stride=False),
- stage_with_dcn=(False, True, True, True)),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_mdconv_c3-c5_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/dcn/faster_rcnn_mdpool_r50_fpn_1x.py b/configs/dcn/faster_rcnn_mdpool_r50_fpn_1x.py
deleted file mode 100644
index c36e65c499314fc787336749a1d6e5602906ee7b..0000000000000000000000000000000000000000
--- a/configs/dcn/faster_rcnn_mdpool_r50_fpn_1x.py
+++ /dev/null
@@ -1,182 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(
- type='ModulatedDeformRoIPoolingPack',
- out_size=7,
- out_channels=256,
- no_trans=False,
- group_size=1,
- trans_std=0.1),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_mdpool_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/dcn/faster_rcnn_r50_fpn_dconv_c3-c5_1x_coco.py b/configs/dcn/faster_rcnn_r50_fpn_dconv_c3-c5_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..08f07fb62e7ec091663d92a3634179a6d6b02aa0
--- /dev/null
+++ b/configs/dcn/faster_rcnn_r50_fpn_dconv_c3-c5_1x_coco.py
@@ -0,0 +1,6 @@
+_base_ = '../faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ backbone=dict(
+ dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
+ stage_with_dcn=(False, True, True, True)))
+work_dir = './work_dirs/faster_rcnn_dconv_c3-c5_r50_fpn_1x'
diff --git a/configs/dcn/faster_rcnn_r50_fpn_dpool_1x_coco.py b/configs/dcn/faster_rcnn_r50_fpn_dpool_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..5a2ed70238a43fcfd056d4d8a3e7d12d8ac0197f
--- /dev/null
+++ b/configs/dcn/faster_rcnn_r50_fpn_dpool_1x_coco.py
@@ -0,0 +1,15 @@
+_base_ = '../faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ bbox_roi_extractor=dict(
+ type='SingleRoIExtractor',
+ roi_layer=dict(
+ _delete_=True,
+ type='DeformRoIPoolingPack',
+ out_size=7,
+ out_channels=256,
+ no_trans=False,
+ group_size=1,
+ trans_std=0.1),
+ out_channels=256,
+ featmap_strides=[4, 8, 16, 32]))
+work_dir = './work_dirs/faster_rcnn_dpool_r50_fpn_1x'
diff --git a/configs/dcn/faster_rcnn_r50_fpn_mdconv_c3-c5_1x_coco.py b/configs/dcn/faster_rcnn_r50_fpn_mdconv_c3-c5_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..7a42f0803af23be25f54a7df799f0be3914f91d6
--- /dev/null
+++ b/configs/dcn/faster_rcnn_r50_fpn_mdconv_c3-c5_1x_coco.py
@@ -0,0 +1,6 @@
+_base_ = '../faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ backbone=dict(
+ dcn=dict(type='DCNv2', deformable_groups=1, fallback_on_stride=False),
+ stage_with_dcn=(False, True, True, True)))
+work_dir = './work_dirs/faster_rcnn_mdconv_c3-c5_r50_fpn_1x'
diff --git a/configs/dcn/faster_rcnn_r50_fpn_mdconv_c3-c5_group4_1x_coco.py b/configs/dcn/faster_rcnn_r50_fpn_mdconv_c3-c5_group4_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..2e22302aefbee5b40c80dfbbb90caeeb1c49c5ab
--- /dev/null
+++ b/configs/dcn/faster_rcnn_r50_fpn_mdconv_c3-c5_group4_1x_coco.py
@@ -0,0 +1,6 @@
+_base_ = '../faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ backbone=dict(
+ dcn=dict(type='DCNv2', deformable_groups=4, fallback_on_stride=False),
+ stage_with_dcn=(False, True, True, True)))
+work_dir = './work_dirs/faster_rcnn_mdconv_c3-c5_group4_r50_fpn_1x'
diff --git a/configs/dcn/faster_rcnn_r50_fpn_mdpool_1x_coco.py b/configs/dcn/faster_rcnn_r50_fpn_mdpool_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..2bf1bf68d4da350953107128f71782e7ef5d4773
--- /dev/null
+++ b/configs/dcn/faster_rcnn_r50_fpn_mdpool_1x_coco.py
@@ -0,0 +1,15 @@
+_base_ = '../faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ bbox_roi_extractor=dict(
+ type='SingleRoIExtractor',
+ roi_layer=dict(
+ _delete_=True,
+ type='ModulatedDeformRoIPoolingPack',
+ out_size=7,
+ out_channels=256,
+ no_trans=False,
+ group_size=1,
+ trans_std=0.1),
+ out_channels=256,
+ featmap_strides=[4, 8, 16, 32]))
+work_dir = './work_dirs/faster_rcnn_mdpool_r50_fpn_1x'
diff --git a/configs/dcn/faster_rcnn_x101_32x4d_fpn_dconv_c3-c5_1x_coco.py b/configs/dcn/faster_rcnn_x101_32x4d_fpn_dconv_c3-c5_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..b09dc5c0e622585aeb2c373503ab9f48ba087335
--- /dev/null
+++ b/configs/dcn/faster_rcnn_x101_32x4d_fpn_dconv_c3-c5_1x_coco.py
@@ -0,0 +1,16 @@
+_base_ = '../faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_32x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=32,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch',
+ dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
+ stage_with_dcn=(False, True, True, True)))
+work_dir = './work_dirs/faster_rcnn_dconv_c3-c5_x101_32x4d_fpn_1x'
diff --git a/configs/dcn/mask_rcnn_dconv_c3-c5_r50_fpn_1x.py b/configs/dcn/mask_rcnn_dconv_c3-c5_r50_fpn_1x.py
deleted file mode 100644
index 802f4e79049f39265f3af4fc997e1fc378e994cd..0000000000000000000000000000000000000000
--- a/configs/dcn/mask_rcnn_dconv_c3-c5_r50_fpn_1x.py
+++ /dev/null
@@ -1,192 +0,0 @@
-# model settings
-model = dict(
- type='MaskRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
- stage_with_dcn=(False, True, True, True)),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_dconv_c3-c5_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/dcn/mask_rcnn_mdconv_c3-c5_r50_fpn_1x.py b/configs/dcn/mask_rcnn_mdconv_c3-c5_r50_fpn_1x.py
deleted file mode 100644
index 1f2c703b3b174f1ac738da22468ae2ea7f70f232..0000000000000000000000000000000000000000
--- a/configs/dcn/mask_rcnn_mdconv_c3-c5_r50_fpn_1x.py
+++ /dev/null
@@ -1,192 +0,0 @@
-# model settings
-model = dict(
- type='MaskRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- dcn=dict(type='DCNv2', deformable_groups=1, fallback_on_stride=False),
- stage_with_dcn=(False, True, True, True)),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_dconv_c3-c5_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/dcn/mask_rcnn_r50_fpn_dconv_c3-c5_1x_coco.py b/configs/dcn/mask_rcnn_r50_fpn_dconv_c3-c5_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..1f7b9a87e3031c824dab02bf18e319c4b308c79d
--- /dev/null
+++ b/configs/dcn/mask_rcnn_r50_fpn_dconv_c3-c5_1x_coco.py
@@ -0,0 +1,6 @@
+_base_ = '../mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ backbone=dict(
+ dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
+ stage_with_dcn=(False, True, True, True)))
+work_dir = './work_dirs/mask_rcnn_dconv_c3-c5_r50_fpn_1x'
diff --git a/configs/dcn/mask_rcnn_r50_fpn_mdconv_c3-c5_1x_coco.py b/configs/dcn/mask_rcnn_r50_fpn_mdconv_c3-c5_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..8518cebf7a68f39c40a6801a445d2c2b688ee368
--- /dev/null
+++ b/configs/dcn/mask_rcnn_r50_fpn_mdconv_c3-c5_1x_coco.py
@@ -0,0 +1,6 @@
+_base_ = '../mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ backbone=dict(
+ dcn=dict(type='DCNv2', deformable_groups=1, fallback_on_stride=False),
+ stage_with_dcn=(False, True, True, True)))
+work_dir = './work_dirs/mask_rcnn_dconv_c3-c5_r50_fpn_1x'
diff --git a/configs/double_heads/dh_faster_rcnn_r50_fpn_1x.py b/configs/double_heads/dh_faster_rcnn_r50_fpn_1x.py
deleted file mode 100644
index 0f3143687373be9636020888ceb9a5f0549f4812..0000000000000000000000000000000000000000
--- a/configs/double_heads/dh_faster_rcnn_r50_fpn_1x.py
+++ /dev/null
@@ -1,179 +0,0 @@
-# model settings
-model = dict(
- type='DoubleHeadRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- reg_roi_scale_factor=1.3,
- bbox_head=dict(
- type='DoubleConvFCBBoxHead',
- num_convs=4,
- num_fcs=2,
- in_channels=256,
- conv_out_channels=1024,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=2.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=2.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/dh_faster_rcnn_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/double_heads/dh_faster_rcnn_r50_fpn_1x_coco.py b/configs/double_heads/dh_faster_rcnn_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..6f3fa49e6cf8eec005bbbe0d864e6db9857c4217
--- /dev/null
+++ b/configs/double_heads/dh_faster_rcnn_r50_fpn_1x_coco.py
@@ -0,0 +1,21 @@
+_base_ = '../faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ type='DoubleHeadRCNN',
+ reg_roi_scale_factor=1.3,
+ bbox_head=dict(
+ _delete_=True,
+ type='DoubleConvFCBBoxHead',
+ num_convs=4,
+ num_fcs=2,
+ in_channels=256,
+ conv_out_channels=1024,
+ fc_out_channels=1024,
+ roi_feat_size=7,
+ num_classes=81,
+ target_means=[0., 0., 0., 0.],
+ target_stds=[0.1, 0.1, 0.2, 0.2],
+ reg_class_agnostic=False,
+ loss_cls=dict(
+ type='CrossEntropyLoss', use_sigmoid=False, loss_weight=2.0),
+ loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=2.0)))
+work_dir = './work_dirs/dh_faster_rcnn_r50_fpn_1x'
diff --git a/configs/empirical_attention/faster_rcnn_r50_fpn_attention_0010_1x.py b/configs/empirical_attention/faster_rcnn_r50_fpn_attention_0010_1x.py
deleted file mode 100644
index 9e196b24d28886e82cf1152d63eab6f00f651d50..0000000000000000000000000000000000000000
--- a/configs/empirical_attention/faster_rcnn_r50_fpn_attention_0010_1x.py
+++ /dev/null
@@ -1,180 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- gen_attention=dict(
- spatial_range=-1, num_heads=8, attention_type='0010', kv_stride=2),
- stage_with_gen_attention=[[], [], [0, 1, 2, 3, 4, 5], [0, 1, 2]],
- ),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_r50_fpn_attention_0010_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/empirical_attention/faster_rcnn_r50_fpn_attention_0010_1x_coco.py b/configs/empirical_attention/faster_rcnn_r50_fpn_attention_0010_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..5f618301c47d54b5f2e07135a21100c98dc94762
--- /dev/null
+++ b/configs/empirical_attention/faster_rcnn_r50_fpn_attention_0010_1x_coco.py
@@ -0,0 +1,8 @@
+_base_ = '../faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ backbone=dict(
+ gen_attention=dict(
+ spatial_range=-1, num_heads=8, attention_type='0010', kv_stride=2),
+ stage_with_gen_attention=[[], [], [0, 1, 2, 3, 4, 5], [0, 1, 2]],
+ ))
+work_dir = './work_dirs/faster_rcnn_r50_fpn_attention_0010_1x'
diff --git a/configs/empirical_attention/faster_rcnn_r50_fpn_attention_0010_dcn_1x.py b/configs/empirical_attention/faster_rcnn_r50_fpn_attention_0010_dcn_1x.py
deleted file mode 100644
index 51827b404947446732b36899d2a86dc813a33a5a..0000000000000000000000000000000000000000
--- a/configs/empirical_attention/faster_rcnn_r50_fpn_attention_0010_dcn_1x.py
+++ /dev/null
@@ -1,182 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- gen_attention=dict(
- spatial_range=-1, num_heads=8, attention_type='0010', kv_stride=2),
- stage_with_gen_attention=[[], [], [0, 1, 2, 3, 4, 5], [0, 1, 2]],
- dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
- stage_with_dcn=(False, True, True, True),
- ),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_r50_fpn_attention_0010_dcn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/empirical_attention/faster_rcnn_r50_fpn_attention_0010_dcn_1x_coco.py b/configs/empirical_attention/faster_rcnn_r50_fpn_attention_0010_dcn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..ed54590a1090897617554ed2f1c3f64b4ad5ea7e
--- /dev/null
+++ b/configs/empirical_attention/faster_rcnn_r50_fpn_attention_0010_dcn_1x_coco.py
@@ -0,0 +1,10 @@
+_base_ = '../faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ backbone=dict(
+ gen_attention=dict(
+ spatial_range=-1, num_heads=8, attention_type='0010', kv_stride=2),
+ stage_with_gen_attention=[[], [], [0, 1, 2, 3, 4, 5], [0, 1, 2]],
+ dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
+ stage_with_dcn=(False, True, True, True),
+ ))
+work_dir = './work_dirs/faster_rcnn_r50_fpn_attention_0010_dcn_1x'
diff --git a/configs/empirical_attention/faster_rcnn_r50_fpn_attention_1111_1x.py b/configs/empirical_attention/faster_rcnn_r50_fpn_attention_1111_1x.py
deleted file mode 100644
index d76d599b2c40703a4857ddb0704eb9ce9395281a..0000000000000000000000000000000000000000
--- a/configs/empirical_attention/faster_rcnn_r50_fpn_attention_1111_1x.py
+++ /dev/null
@@ -1,180 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- gen_attention=dict(
- spatial_range=-1, num_heads=8, attention_type='1111', kv_stride=2),
- stage_with_gen_attention=[[], [], [0, 1, 2, 3, 4, 5], [0, 1, 2]],
- ),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_r50_fpn_attention_1111_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/empirical_attention/faster_rcnn_r50_fpn_attention_1111_1x_coco.py b/configs/empirical_attention/faster_rcnn_r50_fpn_attention_1111_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..62b5b05126e34a1bb6562602342429c6ade2bdaf
--- /dev/null
+++ b/configs/empirical_attention/faster_rcnn_r50_fpn_attention_1111_1x_coco.py
@@ -0,0 +1,8 @@
+_base_ = '../faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ backbone=dict(
+ gen_attention=dict(
+ spatial_range=-1, num_heads=8, attention_type='1111', kv_stride=2),
+ stage_with_gen_attention=[[], [], [0, 1, 2, 3, 4, 5], [0, 1, 2]],
+ ))
+work_dir = './work_dirs/faster_rcnn_r50_fpn_attention_1111_1x'
diff --git a/configs/empirical_attention/faster_rcnn_r50_fpn_attention_1111_dcn_1x.py b/configs/empirical_attention/faster_rcnn_r50_fpn_attention_1111_dcn_1x.py
deleted file mode 100644
index 8132e696100cebae03b2ae554f3ec57ad47fb206..0000000000000000000000000000000000000000
--- a/configs/empirical_attention/faster_rcnn_r50_fpn_attention_1111_dcn_1x.py
+++ /dev/null
@@ -1,182 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- gen_attention=dict(
- spatial_range=-1, num_heads=8, attention_type='1111', kv_stride=2),
- stage_with_gen_attention=[[], [], [0, 1, 2, 3, 4, 5], [0, 1, 2]],
- dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
- stage_with_dcn=(False, True, True, True),
- ),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_r50_fpn_attention_1111_dcn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/empirical_attention/faster_rcnn_r50_fpn_attention_1111_dcn_1x_coco.py b/configs/empirical_attention/faster_rcnn_r50_fpn_attention_1111_dcn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..7c27530977e31ec18d926e3f54efbf618eb0123f
--- /dev/null
+++ b/configs/empirical_attention/faster_rcnn_r50_fpn_attention_1111_dcn_1x_coco.py
@@ -0,0 +1,10 @@
+_base_ = '../faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ backbone=dict(
+ gen_attention=dict(
+ spatial_range=-1, num_heads=8, attention_type='1111', kv_stride=2),
+ stage_with_gen_attention=[[], [], [0, 1, 2, 3, 4, 5], [0, 1, 2]],
+ dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
+ stage_with_dcn=(False, True, True, True),
+ ))
+work_dir = './work_dirs/faster_rcnn_r50_fpn_attention_1111_dcn_1x'
diff --git a/configs/fast_mask_rcnn_r101_fpn_1x.py b/configs/fast_mask_rcnn_r101_fpn_1x.py
deleted file mode 100644
index 8f4079bc144a8f659453d5898383be948d923e63..0000000000000000000000000000000000000000
--- a/configs/fast_mask_rcnn_r101_fpn_1x.py
+++ /dev/null
@@ -1,159 +0,0 @@
-# model settings
-model = dict(
- type='FastRCNN',
- pretrained='torchvision://resnet101',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadProposals', num_max_proposals=2000),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(
- type='Collect',
- keys=['img', 'proposals', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadProposals', num_max_proposals=None),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='ToTensor', keys=['proposals']),
- dict(
- type='ToDataContainer',
- fields=[dict(key='proposals', stack=False)]),
- dict(type='Collect', keys=['img', 'proposals']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- proposal_file=data_root + 'proposals/rpn_r50_fpn_1x_train2017.pkl',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- proposal_file=data_root + 'proposals/rpn_r50_fpn_1x_val2017.pkl',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- proposal_file=data_root + 'proposals/rpn_r50_fpn_1x_val2017.pkl',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/fast_mask_rcnn_r101_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/fast_mask_rcnn_r50_caffe_c4_1x.py b/configs/fast_mask_rcnn_r50_caffe_c4_1x.py
deleted file mode 100644
index 5d32f0e78d2325a4fc92b4ee735ac12bb95043ce..0000000000000000000000000000000000000000
--- a/configs/fast_mask_rcnn_r50_caffe_c4_1x.py
+++ /dev/null
@@ -1,157 +0,0 @@
-# model settings
-norm_cfg = dict(type='BN', requires_grad=False)
-model = dict(
- type='FastRCNN',
- pretrained='open-mmlab://resnet50_caffe',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=3,
- strides=(1, 2, 2),
- dilations=(1, 1, 1),
- out_indices=(2, ),
- frozen_stages=1,
- norm_cfg=norm_cfg,
- norm_eval=True,
- style='caffe'),
- shared_head=dict(
- type='ResLayer',
- depth=50,
- stage=3,
- stride=2,
- dilation=1,
- style='caffe',
- norm_cfg=norm_cfg,
- norm_eval=True),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=1024,
- featmap_strides=[16]),
- bbox_head=dict(
- type='BBoxHead',
- with_avg_pool=True,
- roi_feat_size=7,
- in_channels=2048,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False),
- mask_roi_extractor=None,
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=0,
- in_channels=2048,
- conv_out_channels=256,
- num_classes=81))
-# model training and testing settings
-train_cfg = dict(
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=14,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadProposals', num_max_proposals=2000),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(
- type='Collect',
- keys=['img', 'proposals', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadProposals', num_max_proposals=None),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='ToTensor', keys=['proposals']),
- dict(
- type='ToDataContainer',
- fields=[dict(key='proposals', stack=False)]),
- dict(type='Collect', keys=['img', 'proposals']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- proposal_file=data_root + 'proposals/rpn_r50_c4_1x_train2017.pkl',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- proposal_file=data_root + 'proposals/rpn_r50_c4_1x_val2017.pkl',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- proposal_file=data_root + 'proposals/rpn_r50_c4_1x_val2017.pkl',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/fast_mask_rcnn_r50_caffe_c4_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/fast_mask_rcnn_r50_fpn_1x.py b/configs/fast_mask_rcnn_r50_fpn_1x.py
deleted file mode 100644
index 315fa2c0216eace74b197f423decf45317db6c41..0000000000000000000000000000000000000000
--- a/configs/fast_mask_rcnn_r50_fpn_1x.py
+++ /dev/null
@@ -1,159 +0,0 @@
-# model settings
-model = dict(
- type='FastRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadProposals', num_max_proposals=2000),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(
- type='Collect',
- keys=['img', 'proposals', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadProposals', num_max_proposals=None),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='ToTensor', keys=['proposals']),
- dict(
- type='ToDataContainer',
- fields=[dict(key='proposals', stack=False)]),
- dict(type='Collect', keys=['img', 'proposals']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- proposal_file=data_root + 'proposals/rpn_r50_fpn_1x_train2017.pkl',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- proposal_file=data_root + 'proposals/rpn_r50_fpn_1x_val2017.pkl',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- proposal_file=data_root + 'proposals/rpn_r50_fpn_1x_val2017.pkl',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/fast_mask_rcnn_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/fast_rcnn/fast_rcnn_r50_fpn_1x_coco.py b/configs/fast_rcnn/fast_rcnn_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..db6ea5dd551043ac7d3cd047eaf475bd5d06f56a
--- /dev/null
+++ b/configs/fast_rcnn/fast_rcnn_r50_fpn_1x_coco.py
@@ -0,0 +1,53 @@
+_base_ = [
+ '../_base_/models/fast_rcnn_r50_fpn.py',
+ '../_base_/datasets/coco_detection.py',
+ '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
+]
+dataset_type = 'CocoDataset'
+data_root = 'data/coco/'
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadProposals', num_max_proposals=2000),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'proposals', 'gt_bboxes', 'gt_labels']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadProposals', num_max_proposals=None),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(1333, 800),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='ToTensor', keys=['proposals']),
+ dict(
+ type='ToDataContainer',
+ fields=[dict(key='proposals', stack=False)]),
+ dict(type='Collect', keys=['img', 'proposals']),
+ ])
+]
+data = dict(
+ imgs_per_gpu=2,
+ workers_per_gpu=2,
+ train=dict(
+ proposal_file=data_root + 'proposals/rpn_r50_fpn_1x_train2017.pkl',
+ pipeline=train_pipeline),
+ val=dict(
+ proposal_file=data_root + 'proposals/rpn_r50_fpn_1x_val2017.pkl',
+ pipeline=test_pipeline),
+ test=dict(
+ proposal_file=data_root + 'proposals/rpn_r50_fpn_1x_val2017.pkl',
+ pipeline=test_pipeline))
+work_dir = './work_dirs/fast_rcnn_r50_fpn_1x'
diff --git a/configs/fast_rcnn_r101_fpn_1x.py b/configs/fast_rcnn_r101_fpn_1x.py
deleted file mode 100644
index e46089320d844cb33cfdadddf4765680c861be51..0000000000000000000000000000000000000000
--- a/configs/fast_rcnn_r101_fpn_1x.py
+++ /dev/null
@@ -1,140 +0,0 @@
-# model settings
-model = dict(
- type='FastRCNN',
- pretrained='torchvision://resnet101',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadProposals', num_max_proposals=2000),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'proposals', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadProposals', num_max_proposals=None),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='ToTensor', keys=['proposals']),
- dict(
- type='ToDataContainer',
- fields=[dict(key='proposals', stack=False)]),
- dict(type='Collect', keys=['img', 'proposals']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- proposal_file=data_root + 'proposals/rpn_r50_fpn_1x_train2017.pkl',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- proposal_file=data_root + 'proposals/rpn_r50_fpn_1x_val2017.pkl',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- proposal_file=data_root + 'proposals/rpn_r50_fpn_1x_val2017.pkl',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/fast_rcnn_r101_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/fast_rcnn_r50_caffe_c4_1x.py b/configs/fast_rcnn_r50_caffe_c4_1x.py
deleted file mode 100644
index dec822a383f141020a2f32bcdf7c9f7476eff652..0000000000000000000000000000000000000000
--- a/configs/fast_rcnn_r50_caffe_c4_1x.py
+++ /dev/null
@@ -1,147 +0,0 @@
-# model settings
-norm_cfg = dict(type='BN', requires_grad=False)
-model = dict(
- type='FastRCNN',
- pretrained='open-mmlab://resnet50_caffe',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=3,
- strides=(1, 2, 2),
- dilations=(1, 1, 1),
- out_indices=(2, ),
- frozen_stages=1,
- norm_cfg=norm_cfg,
- norm_eval=True,
- style='caffe'),
- shared_head=dict(
- type='ResLayer',
- depth=50,
- stage=3,
- stride=2,
- dilation=1,
- style='caffe',
- norm_cfg=norm_cfg,
- norm_eval=True),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=1024,
- featmap_strides=[16]),
- bbox_head=dict(
- type='BBoxHead',
- with_avg_pool=True,
- roi_feat_size=7,
- in_channels=2048,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadProposals', num_max_proposals=2000),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'proposals', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadProposals', num_max_proposals=None),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='ToTensor', keys=['proposals']),
- dict(
- type='ToDataContainer',
- fields=[dict(key='proposals', stack=False)]),
- dict(type='Collect', keys=['img', 'proposals']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- proposal_file=data_root + 'proposals/rpn_r50_c4_1x_train2017.pkl',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- proposal_file=data_root + 'proposals/rpn_r50_c4_1x_val2017.pkl',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- proposal_file=data_root + 'proposals/rpn_r50_c4_1x_val2017.pkl',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/fast_rcnn_r50_caffe_c4_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/fast_rcnn_r50_fpn_1x.py b/configs/fast_rcnn_r50_fpn_1x.py
deleted file mode 100644
index dbeab714fb661906e917bb0be83536926049e418..0000000000000000000000000000000000000000
--- a/configs/fast_rcnn_r50_fpn_1x.py
+++ /dev/null
@@ -1,140 +0,0 @@
-# model settings
-model = dict(
- type='FastRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadProposals', num_max_proposals=2000),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'proposals', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadProposals', num_max_proposals=None),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='ToTensor', keys=['proposals']),
- dict(
- type='ToDataContainer',
- fields=[dict(key='proposals', stack=False)]),
- dict(type='Collect', keys=['img', 'proposals']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- proposal_file=data_root + 'proposals/rpn_r50_fpn_1x_train2017.pkl',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- proposal_file=data_root + 'proposals/rpn_r50_fpn_1x_val2017.pkl',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- proposal_file=data_root + 'proposals/rpn_r50_fpn_1x_val2017.pkl',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/fast_rcnn_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/faster_rcnn/faster_rcnn_r101_fpn_1x_coco.py b/configs/faster_rcnn/faster_rcnn_r101_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..c0b51c59d11b2a6c388ce62c3f62e548fed0c035
--- /dev/null
+++ b/configs/faster_rcnn/faster_rcnn_r101_fpn_1x_coco.py
@@ -0,0 +1,3 @@
+_base_ = './faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(pretrained='torchvision://resnet101', backbone=dict(depth=101))
+work_dir = './work_dirs/faster_rcnn_r101_fpn_1x'
diff --git a/configs/faster_rcnn/faster_rcnn_r50_caffe_c4_1x_coco.py b/configs/faster_rcnn/faster_rcnn_r50_caffe_c4_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..5a5b7ccd8cd968e433f6990238f5a71792a600f4
--- /dev/null
+++ b/configs/faster_rcnn/faster_rcnn_r50_caffe_c4_1x_coco.py
@@ -0,0 +1,40 @@
+_base_ = [
+ '../_base_/models/faster_rcnn_r50_caffe_c4.py',
+ '../_base_/datasets/coco_detection.py',
+ '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
+]
+# use caffe img_norm
+img_norm_cfg = dict(
+ mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(1333, 800),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ train=dict(pipeline=train_pipeline),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+# optimizer
+optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
+work_dir = './work_dirs/faster_rcnn_r50_caffe_c4_1x'
diff --git a/configs/faster_rcnn/faster_rcnn_r50_caffe_fpn_1x_coco.py b/configs/faster_rcnn/faster_rcnn_r50_caffe_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..8fdc75b4d1a61477bb1066cd6fcd08d1fb99da5f
--- /dev/null
+++ b/configs/faster_rcnn/faster_rcnn_r50_caffe_fpn_1x_coco.py
@@ -0,0 +1,37 @@
+_base_ = './faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnet50_caffe',
+ backbone=dict(norm_cfg=dict(requires_grad=False), style='caffe'))
+# use caffe img_norm
+img_norm_cfg = dict(
+ mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(1333, 800),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ train=dict(pipeline=train_pipeline),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+work_dir = './work_dirs/faster_rcnn_r50_caffe_fpn_1x'
diff --git a/configs/faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py b/configs/faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..4df206625ee54bddfafd354d5c9bca05711334aa
--- /dev/null
+++ b/configs/faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py
@@ -0,0 +1,6 @@
+_base_ = [
+ '../_base_/models/faster_rcnn_r50_fpn.py',
+ '../_base_/datasets/coco_detection.py',
+ '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
+]
+work_dir = './work_dirs/faster_rcnn_r50_fpn_1x'
diff --git a/configs/faster_rcnn/faster_rcnn_r50_fpn_ohem_1x_coco.py b/configs/faster_rcnn/faster_rcnn_r50_fpn_ohem_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..0a02fdf994ec70c48e4a206200d5668d8b900b1d
--- /dev/null
+++ b/configs/faster_rcnn/faster_rcnn_r50_fpn_ohem_1x_coco.py
@@ -0,0 +1,3 @@
+_base_ = './faster_rcnn_r50_fpn_1x_coco.py'
+train_cfg = dict(rcnn=dict(sampler=dict(type='OHEMSampler')))
+work_dir = './work_dirs/faster_rcnn_r50_fpn_1x'
diff --git a/configs/faster_rcnn/faster_rcnn_x101_32x4d_fpn_1x_coco.py b/configs/faster_rcnn/faster_rcnn_x101_32x4d_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..2dc66a86ed847d4b61018f24ffe50ae93fa19fff
--- /dev/null
+++ b/configs/faster_rcnn/faster_rcnn_x101_32x4d_fpn_1x_coco.py
@@ -0,0 +1,14 @@
+_base_ = './faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_32x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=32,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+work_dir = './work_dirs/faster_rcnn_x101_32x4d_fpn_1x'
diff --git a/configs/faster_rcnn/faster_rcnn_x101_64x4d_fpn_1x_coco.py b/configs/faster_rcnn/faster_rcnn_x101_64x4d_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..0e85ee5ea5db6b01267a8f175a37ca8ef15c85fb
--- /dev/null
+++ b/configs/faster_rcnn/faster_rcnn_x101_64x4d_fpn_1x_coco.py
@@ -0,0 +1,14 @@
+_base_ = './faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_64x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=64,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+work_dir = './work_dirs/faster_rcnn_x101_64x4d_fpn_1x'
diff --git a/configs/faster_rcnn_ohem_r50_fpn_1x.py b/configs/faster_rcnn_ohem_r50_fpn_1x.py
deleted file mode 100644
index baad9c85736d66d60173a24d0ce59a59b589e9f8..0000000000000000000000000000000000000000
--- a/configs/faster_rcnn_ohem_r50_fpn_1x.py
+++ /dev/null
@@ -1,176 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='OHEMSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/faster_rcnn_r101_fpn_1x.py b/configs/faster_rcnn_r101_fpn_1x.py
deleted file mode 100644
index d3735e5fdfb6cb055ae66c7d1e70d1565ce4d755..0000000000000000000000000000000000000000
--- a/configs/faster_rcnn_r101_fpn_1x.py
+++ /dev/null
@@ -1,176 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='torchvision://resnet101',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_r101_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/faster_rcnn_x101_32x4d_fpn_1x.py b/configs/faster_rcnn_x101_32x4d_fpn_1x.py
deleted file mode 100644
index 9e14dc9f5ea4d3ceeb07a9e3564ffbca6c5a3657..0000000000000000000000000000000000000000
--- a/configs/faster_rcnn_x101_32x4d_fpn_1x.py
+++ /dev/null
@@ -1,178 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='open-mmlab://resnext101_32x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=32,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_x101_32x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/faster_rcnn_x101_64x4d_fpn_1x.py b/configs/faster_rcnn_x101_64x4d_fpn_1x.py
deleted file mode 100644
index 5454cfe170583d310352798c1eb0ea6c67a2b1a9..0000000000000000000000000000000000000000
--- a/configs/faster_rcnn_x101_64x4d_fpn_1x.py
+++ /dev/null
@@ -1,178 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='open-mmlab://resnext101_64x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=64,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_x101_64x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/fcos/fcos_center_r50_caffe_fpn_gn-head_4x4_1x_coco.py b/configs/fcos/fcos_center_r50_caffe_fpn_gn-head_4x4_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..d655f854aa70217106addd33d238ff172919d5c3
--- /dev/null
+++ b/configs/fcos/fcos_center_r50_caffe_fpn_gn-head_4x4_1x_coco.py
@@ -0,0 +1,3 @@
+_base_ = './fcos_r50_caffe_fpn_gn-head_4x4_1x_coco.py'
+model = dict(bbox_head=dict(center_sampling=True, center_sample_radius=1.5))
+work_dir = './work_dirs/fcos_center_r50_caffe_fpn_gn_1x_4gpu'
diff --git a/configs/fcos/fcos_center_r50_caffe_fpn_gn_1x_4gpu.py.py b/configs/fcos/fcos_center_r50_caffe_fpn_gn_1x_4gpu.py.py
deleted file mode 100644
index 678fb0a531da55b89bcc8ef1b6a323888ca64759..0000000000000000000000000000000000000000
--- a/configs/fcos/fcos_center_r50_caffe_fpn_gn_1x_4gpu.py.py
+++ /dev/null
@@ -1,137 +0,0 @@
-# model settings
-model = dict(
- type='FCOS',
- pretrained='open-mmlab://resnet50_caffe',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=False),
- style='caffe'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- extra_convs_on_inputs=False, # use P5
- num_outs=5,
- relu_before_extra_convs=True),
- bbox_head=dict(
- type='FCOSHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- strides=[8, 16, 32, 64, 128],
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox=dict(type='IoULoss', loss_weight=1.0),
- loss_centerness=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- center_sampling=True,
- center_sample_radius=1.5))
-# training and testing settings
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=4,
- workers_per_gpu=4,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(
- type='SGD',
- lr=0.01,
- momentum=0.9,
- weight_decay=0.0001,
- paramwise_options=dict(bias_lr_mult=2., bias_decay_mult=0.))
-optimizer_config = dict(grad_clip=None)
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='constant',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/fcos_center_r50_caffe_fpn_gn_1x_4gpu'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/fcos/fcos_mstrain_640_800_r101_caffe_fpn_gn_2x_4gpu.py b/configs/fcos/fcos_mstrain_640_800_r101_caffe_fpn_gn_2x_4gpu.py
deleted file mode 100644
index 40a358e9445280c0b8627e21cd4b0a8b5c43d8bc..0000000000000000000000000000000000000000
--- a/configs/fcos/fcos_mstrain_640_800_r101_caffe_fpn_gn_2x_4gpu.py
+++ /dev/null
@@ -1,139 +0,0 @@
-# model settings
-model = dict(
- type='FCOS',
- pretrained='open-mmlab://resnet101_caffe',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=False),
- style='caffe'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- extra_convs_on_inputs=False, # use P5
- num_outs=5,
- relu_before_extra_convs=True),
- bbox_head=dict(
- type='FCOSHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- strides=[8, 16, 32, 64, 128],
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox=dict(type='IoULoss', loss_weight=1.0),
- loss_centerness=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(
- type='Resize',
- img_scale=[(1333, 640), (1333, 800)],
- multiscale_mode='value',
- keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=4,
- workers_per_gpu=4,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(
- type='SGD',
- lr=0.01,
- momentum=0.9,
- weight_decay=0.0001,
- paramwise_options=dict(bias_lr_mult=2., bias_decay_mult=0.))
-optimizer_config = dict(grad_clip=None)
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='constant',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/fcos_mstrain_640_800_r101_caffe_fpn_gn_2x_4gpu'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/fcos/fcos_mstrain_640_800_x101_64x4d_fpn_gn_2x.py b/configs/fcos/fcos_mstrain_640_800_x101_64x4d_fpn_gn_2x.py
deleted file mode 100644
index 3ba6505024b346514c6f8fe3913b81906b9bf073..0000000000000000000000000000000000000000
--- a/configs/fcos/fcos_mstrain_640_800_x101_64x4d_fpn_gn_2x.py
+++ /dev/null
@@ -1,141 +0,0 @@
-# model settings
-model = dict(
- type='FCOS',
- pretrained='open-mmlab://resnext101_64x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=64,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- extra_convs_on_inputs=False, # use P5
- num_outs=5,
- relu_before_extra_convs=True),
- bbox_head=dict(
- type='FCOSHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- strides=[8, 16, 32, 64, 128],
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox=dict(type='IoULoss', loss_weight=1.0),
- loss_centerness=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(
- type='Resize',
- img_scale=[(1333, 640), (1333, 800)],
- multiscale_mode='value',
- keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(
- type='SGD',
- lr=0.01,
- momentum=0.9,
- weight_decay=0.0001,
- paramwise_options=dict(bias_lr_mult=2., bias_decay_mult=0.))
-optimizer_config = dict(grad_clip=None)
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='constant',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/fcos_mstrain_640_800_x101_64x4d_fpn_gn_2x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/fcos/fcos_r101_caffe_fpn_gn-head_mstrain_640-800_4x4_2x_coco.py b/configs/fcos/fcos_r101_caffe_fpn_gn-head_mstrain_640-800_4x4_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..e4795fde3694653dc2be38cc21f2d2477940ea7b
--- /dev/null
+++ b/configs/fcos/fcos_r101_caffe_fpn_gn-head_mstrain_640-800_4x4_2x_coco.py
@@ -0,0 +1,44 @@
+_base_ = './fcos_r50_caffe_fpn_gn-head_4x4_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnet101_caffe', backbone=dict(depth=101))
+img_norm_cfg = dict(
+ mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(
+ type='Resize',
+ img_scale=[(1333, 640), (1333, 800)],
+ multiscale_mode='value',
+ keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(1333, 800),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ imgs_per_gpu=4,
+ workers_per_gpu=4,
+ train=dict(pipeline=train_pipeline),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+# learning policy
+lr_config = dict(step=[16, 22])
+total_epochs = 24
+work_dir = './work_dirs/fcos_mstrain_640_800_r101_caffe_fpn_gn_2x_4gpu'
diff --git a/configs/fcos/fcos_r50_caffe_fpn_gn_1x_4gpu.py b/configs/fcos/fcos_r50_caffe_fpn_gn-head_4x4_1x_coco.py
similarity index 66%
rename from configs/fcos/fcos_r50_caffe_fpn_gn_1x_4gpu.py
rename to configs/fcos/fcos_r50_caffe_fpn_gn-head_4x4_1x_coco.py
index fb37d3149c85b95cd9e218dfd092de6ce348e387..fa19da5278be2f4e321a25c688c7ea65e75dbf2e 100644
--- a/configs/fcos/fcos_r50_caffe_fpn_gn_1x_4gpu.py
+++ b/configs/fcos/fcos_r50_caffe_fpn_gn-head_4x4_1x_coco.py
@@ -1,3 +1,7 @@
+_base_ = [
+ '../_base_/datasets/coco_detection.py',
+ '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
+]
# model settings
model = dict(
type='FCOS',
@@ -52,9 +56,6 @@ test_cfg = dict(
score_thr=0.05,
nms=dict(type='nms', iou_thr=0.5),
max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
img_norm_cfg = dict(
mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
train_pipeline = [
@@ -85,51 +86,14 @@ test_pipeline = [
data = dict(
imgs_per_gpu=4,
workers_per_gpu=4,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
+ train=dict(pipeline=train_pipeline),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
# optimizer
optimizer = dict(
- type='SGD',
- lr=0.01,
- momentum=0.9,
- weight_decay=0.0001,
- paramwise_options=dict(bias_lr_mult=2., bias_decay_mult=0.))
-optimizer_config = dict(grad_clip=None)
+ lr=0.01, paramwise_options=dict(bias_lr_mult=2., bias_decay_mult=0.))
+optimizer_config = dict(_delete_=True, grad_clip=None)
# learning policy
-lr_config = dict(
- policy='step',
- warmup='constant',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
+lr_config = dict(warmup='constant')
total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
work_dir = './work_dirs/fcos_r50_caffe_fpn_gn_1x_4gpu'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/fcos/fcos_x101_64x4d_fpn_gn-head_mstrain_640-800_4x2_2x_coco.py b/configs/fcos/fcos_x101_64x4d_fpn_gn-head_mstrain_640-800_4x2_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..a1fe0d9fd0d9a9988f067ca9eb123afbbf9e6f07
--- /dev/null
+++ b/configs/fcos/fcos_x101_64x4d_fpn_gn-head_mstrain_640-800_4x2_2x_coco.py
@@ -0,0 +1,58 @@
+_base_ = './fcos_r50_caffe_fpn_gn-head_4x4_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_64x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=64,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(
+ type='Resize',
+ img_scale=[(1333, 640), (1333, 800)],
+ multiscale_mode='value',
+ keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(1333, 800),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ imgs_per_gpu=2,
+ workers_per_gpu=2,
+ train=dict(pipeline=train_pipeline),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+# optimizer
+optimizer = dict(
+ lr=0.01, paramwise_options=dict(bias_lr_mult=2., bias_decay_mult=0.))
+optimizer_config = dict(_delete_=True, grad_clip=None)
+# learning policy
+lr_config = dict(warmup='constant', step=[16, 22])
+total_epochs = 24
+work_dir = './work_dirs/fcos_mstrain_640_800_x101_64x4d_fpn_gn_2x'
diff --git a/configs/foveabox/fovea_align_gn_ms_r101_fpn_4gpu_2x.py b/configs/foveabox/fovea_align_gn_ms_r101_fpn_4gpu_2x.py
deleted file mode 100644
index 502ff3ee76705ea23064706eee0df0971d095e11..0000000000000000000000000000000000000000
--- a/configs/foveabox/fovea_align_gn_ms_r101_fpn_4gpu_2x.py
+++ /dev/null
@@ -1,126 +0,0 @@
-# model settings
-model = dict(
- type='FOVEA',
- pretrained='torchvision://resnet101',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- num_outs=5,
- add_extra_convs=True),
- bbox_head=dict(
- type='FoveaHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- strides=[8, 16, 32, 64, 128],
- base_edge_list=[16, 32, 64, 128, 256],
- scale_ranges=((1, 64), (32, 128), (64, 256), (128, 512), (256, 2048)),
- sigma=0.4,
- with_deform=True,
- norm_cfg=dict(type='GN', num_groups=32, requires_grad=True),
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=1.50,
- alpha=0.4,
- loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict()
-test_cfg = dict(
- nms_pre=1000,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(
- type='Resize',
- img_scale=[(1333, 640), (1333, 800)],
- multiscale_mode='value',
- keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=4,
- workers_per_gpu=4,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/fovea_align_gn_ms_r101_fpn_4gpu_2x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/foveabox/fovea_align_gn_ms_r50_fpn_4gpu_2x.py b/configs/foveabox/fovea_align_gn_ms_r50_fpn_4gpu_2x.py
deleted file mode 100644
index a4960660ccf39d35b89a660b95973c239f816880..0000000000000000000000000000000000000000
--- a/configs/foveabox/fovea_align_gn_ms_r50_fpn_4gpu_2x.py
+++ /dev/null
@@ -1,126 +0,0 @@
-# model settings
-model = dict(
- type='FOVEA',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- num_outs=5,
- add_extra_convs=True),
- bbox_head=dict(
- type='FoveaHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- strides=[8, 16, 32, 64, 128],
- base_edge_list=[16, 32, 64, 128, 256],
- scale_ranges=((1, 64), (32, 128), (64, 256), (128, 512), (256, 2048)),
- sigma=0.4,
- with_deform=True,
- norm_cfg=dict(type='GN', num_groups=32, requires_grad=True),
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=1.50,
- alpha=0.4,
- loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict()
-test_cfg = dict(
- nms_pre=1000,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(
- type='Resize',
- img_scale=[(1333, 640), (1333, 800)],
- multiscale_mode='value',
- keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=4,
- workers_per_gpu=4,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/fovea_align_gn_ms_r50_fpn_4gpu_2x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/foveabox/fovea_align_gn_r101_fpn_4gpu_2x.py b/configs/foveabox/fovea_align_gn_r101_fpn_4gpu_2x.py
deleted file mode 100644
index 8b7acaf61e7250e6855d0589b98cb39f1637704f..0000000000000000000000000000000000000000
--- a/configs/foveabox/fovea_align_gn_r101_fpn_4gpu_2x.py
+++ /dev/null
@@ -1,121 +0,0 @@
-# model settings
-model = dict(
- type='FOVEA',
- pretrained='torchvision://resnet101',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- num_outs=5,
- add_extra_convs=True),
- bbox_head=dict(
- type='FoveaHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- strides=[8, 16, 32, 64, 128],
- base_edge_list=[16, 32, 64, 128, 256],
- scale_ranges=((1, 64), (32, 128), (64, 256), (128, 512), (256, 2048)),
- sigma=0.4,
- with_deform=True,
- norm_cfg=dict(type='GN', num_groups=32, requires_grad=True),
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=1.50,
- alpha=0.4,
- loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict()
-test_cfg = dict(
- nms_pre=1000,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=4,
- workers_per_gpu=4,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/fovea_align_gn_r101_fpn_4gpu_2x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/foveabox/fovea_align_gn_r50_fpn_4gpu_2x.py b/configs/foveabox/fovea_align_gn_r50_fpn_4gpu_2x.py
deleted file mode 100644
index 15efc4c5a8738ba8c519cadb624fb1086a7d3ba3..0000000000000000000000000000000000000000
--- a/configs/foveabox/fovea_align_gn_r50_fpn_4gpu_2x.py
+++ /dev/null
@@ -1,121 +0,0 @@
-# model settings
-model = dict(
- type='FOVEA',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- num_outs=5,
- add_extra_convs=True),
- bbox_head=dict(
- type='FoveaHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- strides=[8, 16, 32, 64, 128],
- base_edge_list=[16, 32, 64, 128, 256],
- scale_ranges=((1, 64), (32, 128), (64, 256), (128, 512), (256, 2048)),
- sigma=0.4,
- with_deform=True,
- norm_cfg=dict(type='GN', num_groups=32, requires_grad=True),
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=1.50,
- alpha=0.4,
- loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict()
-test_cfg = dict(
- nms_pre=1000,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=4,
- workers_per_gpu=4,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/fovea_align_gn_r50_fpn_4gpu_2x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/foveabox/fovea_align_r101_fpn_gn-head_4x4_2x_coco.py b/configs/foveabox/fovea_align_r101_fpn_gn-head_4x4_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..e2ea2f2e20dae4482c641e9070747b8944874633
--- /dev/null
+++ b/configs/foveabox/fovea_align_r101_fpn_gn-head_4x4_2x_coco.py
@@ -0,0 +1,11 @@
+_base_ = './fovea_r50_fpn_4x4_1x_coco.py'
+model = dict(
+ pretrained='torchvision://resnet101',
+ backbone=dict(depth=101),
+ bbox_head=dict(
+ with_deform=True,
+ norm_cfg=dict(type='GN', num_groups=32, requires_grad=True)))
+# learning policy
+lr_config = dict(step=[16, 22])
+total_epochs = 24
+work_dir = './work_dirs/fovea_align_gn_r101_fpn_4gpu_2x'
diff --git a/configs/foveabox/fovea_align_r101_fpn_gn-head_mstrain_640-800_4x4_2x_coco.py b/configs/foveabox/fovea_align_r101_fpn_gn-head_mstrain_640-800_4x4_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..7717a0e24e697d61d44930c21ab9ab0d45c3b4e6
--- /dev/null
+++ b/configs/foveabox/fovea_align_r101_fpn_gn-head_mstrain_640-800_4x4_2x_coco.py
@@ -0,0 +1,28 @@
+_base_ = './fovea_r50_fpn_4x4_1x_coco.py'
+model = dict(
+ pretrained='torchvision://resnet101',
+ backbone=dict(depth=101),
+ bbox_head=dict(
+ with_deform=True,
+ norm_cfg=dict(type='GN', num_groups=32, requires_grad=True)))
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(
+ type='Resize',
+ img_scale=[(1333, 640), (1333, 800)],
+ multiscale_mode='value',
+ keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+data = dict(train=dict(pipeline=train_pipeline))
+# learning policy
+lr_config = dict(step=[16, 22])
+total_epochs = 24
+work_dir = './work_dirs/fovea_align_gn_ms_r101_fpn_4gpu_2x'
diff --git a/configs/foveabox/fovea_align_r50_fpn_gn-head_4x4_2x_coco.py b/configs/foveabox/fovea_align_r50_fpn_gn-head_4x4_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..0e35fb5361ab23f0a34e282457417e5a9a97cf4c
--- /dev/null
+++ b/configs/foveabox/fovea_align_r50_fpn_gn-head_4x4_2x_coco.py
@@ -0,0 +1,9 @@
+_base_ = './fovea_r50_fpn_4x4_1x_coco.py'
+model = dict(
+ bbox_head=dict(
+ with_deform=True,
+ norm_cfg=dict(type='GN', num_groups=32, requires_grad=True)))
+# learning policy
+lr_config = dict(step=[16, 22])
+total_epochs = 24
+work_dir = './work_dirs/fovea_align_gn_r50_fpn_4gpu_2x'
diff --git a/configs/foveabox/fovea_align_r50_fpn_gn-head_mstrain_640-800_4x4_2x_coco.py b/configs/foveabox/fovea_align_r50_fpn_gn-head_mstrain_640-800_4x4_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..15068d2b2289e8eadc182eb8bf1bc1ffa1230e00
--- /dev/null
+++ b/configs/foveabox/fovea_align_r50_fpn_gn-head_mstrain_640-800_4x4_2x_coco.py
@@ -0,0 +1,26 @@
+_base_ = './fovea_r50_fpn_4x4_1x_coco.py'
+model = dict(
+ bbox_head=dict(
+ with_deform=True,
+ norm_cfg=dict(type='GN', num_groups=32, requires_grad=True)))
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(
+ type='Resize',
+ img_scale=[(1333, 640), (1333, 800)],
+ multiscale_mode='value',
+ keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+data = dict(train=dict(pipeline=train_pipeline))
+# learning policy
+lr_config = dict(step=[16, 22])
+total_epochs = 24
+work_dir = './work_dirs/fovea_align_gn_ms_r50_fpn_4gpu_2x'
diff --git a/configs/foveabox/fovea_r50_fpn_4gpu_1x.py b/configs/foveabox/fovea_r50_fpn_4gpu_1x.py
deleted file mode 100644
index 724a8aab327b7da75497ad0125919712c332b2d2..0000000000000000000000000000000000000000
--- a/configs/foveabox/fovea_r50_fpn_4gpu_1x.py
+++ /dev/null
@@ -1,121 +0,0 @@
-# model settings
-model = dict(
- type='FOVEA',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- num_outs=5,
- add_extra_convs=True),
- bbox_head=dict(
- type='FoveaHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- strides=[8, 16, 32, 64, 128],
- base_edge_list=[16, 32, 64, 128, 256],
- scale_ranges=((1, 64), (32, 128), (64, 256), (128, 512), (256, 2048)),
- sigma=0.4,
- with_deform=False,
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=1.50,
- alpha=0.4,
- loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict()
-test_cfg = dict(
- nms_pre=1000,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=4,
- workers_per_gpu=4,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/fovea_r50_fpn_4gpu_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/foveabox/fovea_r50_fpn_4x4_1x_coco.py b/configs/foveabox/fovea_r50_fpn_4x4_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..c201cfbfe3ab443340d47d0f46cf6f862c618df1
--- /dev/null
+++ b/configs/foveabox/fovea_r50_fpn_4x4_1x_coco.py
@@ -0,0 +1,52 @@
+_base_ = [
+ '../_base_/datasets/coco_detection.py',
+ '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
+]
+# model settings
+model = dict(
+ type='FOVEA',
+ pretrained='torchvision://resnet50',
+ backbone=dict(
+ type='ResNet',
+ depth=50,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'),
+ neck=dict(
+ type='FPN',
+ in_channels=[256, 512, 1024, 2048],
+ out_channels=256,
+ start_level=1,
+ num_outs=5,
+ add_extra_convs=True),
+ bbox_head=dict(
+ type='FoveaHead',
+ num_classes=81,
+ in_channels=256,
+ stacked_convs=4,
+ feat_channels=256,
+ strides=[8, 16, 32, 64, 128],
+ base_edge_list=[16, 32, 64, 128, 256],
+ scale_ranges=((1, 64), (32, 128), (64, 256), (128, 512), (256, 2048)),
+ sigma=0.4,
+ with_deform=False,
+ loss_cls=dict(
+ type='FocalLoss',
+ use_sigmoid=True,
+ gamma=1.50,
+ alpha=0.4,
+ loss_weight=1.0),
+ loss_bbox=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0)))
+# training and testing settings
+train_cfg = dict()
+test_cfg = dict(
+ nms_pre=1000,
+ score_thr=0.05,
+ nms=dict(type='nms', iou_thr=0.5),
+ max_per_img=100)
+data = dict(imgs_per_gpu=4, workers_per_gpu=4)
+# optimizer
+optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
+work_dir = './work_dirs/fovea_r50_fpn_4gpu_1x'
diff --git a/configs/fp16/faster_rcnn_r50_fpn_fp16_1x.py b/configs/fp16/faster_rcnn_r50_fpn_fp16_1x.py
deleted file mode 100644
index ba3a4db66e2a4603d3bdc17c88b1802bc14ca2df..0000000000000000000000000000000000000000
--- a/configs/fp16/faster_rcnn_r50_fpn_fp16_1x.py
+++ /dev/null
@@ -1,179 +0,0 @@
-# fp16 settings
-fp16 = dict(loss_scale=512.)
-
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_r50_fpn_fp16_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/fp16/faster_rcnn_r50_fpn_fp16_1x_coco.py b/configs/fp16/faster_rcnn_r50_fpn_fp16_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..37196fecff959da3dca88b82df498d28fcff9a9a
--- /dev/null
+++ b/configs/fp16/faster_rcnn_r50_fpn_fp16_1x_coco.py
@@ -0,0 +1,4 @@
+_base_ = '../faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
+# fp16 settings
+fp16 = dict(loss_scale=512.)
+work_dir = './work_dirs/faster_rcnn_r50_fpn_fp16_1x'
diff --git a/configs/fp16/mask_rcnn_r50_fpn_fp16_1x.py b/configs/fp16/mask_rcnn_r50_fpn_fp16_1x.py
deleted file mode 100644
index 34bb1489235d19bae21538d9969ecb6b72268793..0000000000000000000000000000000000000000
--- a/configs/fp16/mask_rcnn_r50_fpn_fp16_1x.py
+++ /dev/null
@@ -1,193 +0,0 @@
-# fp16 settings
-fp16 = dict(loss_scale=512.)
-
-# model settings
-model = dict(
- type='MaskRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_r50_fpn_fp16_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/fp16/mask_rcnn_r50_fpn_fp16_1x_coco.py b/configs/fp16/mask_rcnn_r50_fpn_fp16_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..ba14b08af7dc68c5ee2b735eb7c8bb174697bd9c
--- /dev/null
+++ b/configs/fp16/mask_rcnn_r50_fpn_fp16_1x_coco.py
@@ -0,0 +1,4 @@
+_base_ = '../mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py'
+# fp16 settings
+fp16 = dict(loss_scale=512.)
+work_dir = './work_dirs/mask_rcnn_r50_fpn_fp16_1x'
diff --git a/configs/fp16/retinanet_r50_fpn_fp16_1x.py b/configs/fp16/retinanet_r50_fpn_fp16_1x.py
deleted file mode 100644
index 3eebebb64e953e3e12c380d8ed058be1de08353d..0000000000000000000000000000000000000000
--- a/configs/fp16/retinanet_r50_fpn_fp16_1x.py
+++ /dev/null
@@ -1,134 +0,0 @@
-# fp16 settings
-fp16 = dict(loss_scale=512.)
-
-# model settings
-model = dict(
- type='RetinaNet',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5),
- bbox_head=dict(
- type='RetinaHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- octave_base_scale=4,
- scales_per_octave=3,
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[8, 16, 32, 64, 128],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/retinanet_r50_fpn_fp16_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/fp16/retinanet_r50_fpn_fp16_1x_coco.py b/configs/fp16/retinanet_r50_fpn_fp16_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..70cfb7d8baf3dcd634c102fe924e422d7d2610f2
--- /dev/null
+++ b/configs/fp16/retinanet_r50_fpn_fp16_1x_coco.py
@@ -0,0 +1,4 @@
+_base_ = '../retinanet/retinanet_r50_fpn_1x_coco.py'
+# fp16 settings
+fp16 = dict(loss_scale=512.)
+work_dir = './work_dirs/retinanet_r50_fpn_fp16_1x'
diff --git a/configs/free_anchor/retinanet_free_anchor_r101_fpn_1x.py b/configs/free_anchor/retinanet_free_anchor_r101_fpn_1x.py
deleted file mode 100644
index 134de95153ebb99a84f832eb856be04585635c02..0000000000000000000000000000000000000000
--- a/configs/free_anchor/retinanet_free_anchor_r101_fpn_1x.py
+++ /dev/null
@@ -1,125 +0,0 @@
-# model settings
-model = dict(
- type='RetinaNet',
- pretrained='torchvision://resnet101',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5),
- bbox_head=dict(
- type='FreeAnchorRetinaHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- octave_base_scale=4,
- scales_per_octave=3,
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[8, 16, 32, 64, 128],
- target_means=[.0, .0, .0, .0],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- loss_bbox=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.75)))
-# training and testing settings
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/retinanet_free_anchor_r101_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/free_anchor/retinanet_free_anchor_r101_fpn_1x_coco.py b/configs/free_anchor/retinanet_free_anchor_r101_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..44fcacfe3f0e6335af4e1c9d7bf974c8fce507f5
--- /dev/null
+++ b/configs/free_anchor/retinanet_free_anchor_r101_fpn_1x_coco.py
@@ -0,0 +1,3 @@
+_base_ = './retinanet_free_anchor_r50_fpn_1x_coco.py'
+model = dict(pretrained='torchvision://resnet101', backbone=dict(depth=101))
+work_dir = './work_dirs/retinanet_free_anchor_r101_fpn_1x'
diff --git a/configs/free_anchor/retinanet_free_anchor_r50_fpn_1x.py b/configs/free_anchor/retinanet_free_anchor_r50_fpn_1x.py
deleted file mode 100644
index 93a211a6146fe1725b544af3e4a2b9fb33ede7b6..0000000000000000000000000000000000000000
--- a/configs/free_anchor/retinanet_free_anchor_r50_fpn_1x.py
+++ /dev/null
@@ -1,125 +0,0 @@
-# model settings
-model = dict(
- type='RetinaNet',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5),
- bbox_head=dict(
- type='FreeAnchorRetinaHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- octave_base_scale=4,
- scales_per_octave=3,
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[8, 16, 32, 64, 128],
- target_means=[.0, .0, .0, .0],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- loss_bbox=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.75)))
-# training and testing settings
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/retinanet_free_anchor_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/free_anchor/retinanet_free_anchor_r50_fpn_1x_coco.py b/configs/free_anchor/retinanet_free_anchor_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..ad50440bf3cf277cb86a6a1756a59105dd707a40
--- /dev/null
+++ b/configs/free_anchor/retinanet_free_anchor_r50_fpn_1x_coco.py
@@ -0,0 +1,17 @@
+_base_ = '../retinanet/retinanet_r50_fpn_1x_coco.py'
+model = dict(
+ bbox_head=dict(
+ _delete_=True,
+ type='FreeAnchorRetinaHead',
+ num_classes=81,
+ in_channels=256,
+ stacked_convs=4,
+ feat_channels=256,
+ octave_base_scale=4,
+ scales_per_octave=3,
+ anchor_ratios=[0.5, 1.0, 2.0],
+ anchor_strides=[8, 16, 32, 64, 128],
+ target_means=[.0, .0, .0, .0],
+ target_stds=[0.1, 0.1, 0.2, 0.2],
+ loss_bbox=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.75)))
+work_dir = './work_dirs/retinanet_free_anchor_r50_fpn_1x'
diff --git a/configs/free_anchor/retinanet_free_anchor_x101-32x4d_fpn_1x.py b/configs/free_anchor/retinanet_free_anchor_x101-32x4d_fpn_1x.py
deleted file mode 100644
index 5edec0d29920c8b2b0cab8008961e7aeb77e6782..0000000000000000000000000000000000000000
--- a/configs/free_anchor/retinanet_free_anchor_x101-32x4d_fpn_1x.py
+++ /dev/null
@@ -1,127 +0,0 @@
-# model settings
-model = dict(
- type='RetinaNet',
- pretrained='open-mmlab://resnext101_32x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=32,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5),
- bbox_head=dict(
- type='FreeAnchorRetinaHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- octave_base_scale=4,
- scales_per_octave=3,
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[8, 16, 32, 64, 128],
- target_means=[.0, .0, .0, .0],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- loss_bbox=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.75)))
-# training and testing settings
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/retinanet_free_anchor_x101-32x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/free_anchor/retinanet_free_anchor_x101-32x4d_fpn_1x_coco.py b/configs/free_anchor/retinanet_free_anchor_x101-32x4d_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..c3b607412781387d4c1c287ee93818ec7313abf0
--- /dev/null
+++ b/configs/free_anchor/retinanet_free_anchor_x101-32x4d_fpn_1x_coco.py
@@ -0,0 +1,13 @@
+_base_ = './retinanet_free_anchor_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_32x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=32,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ style='pytorch'))
+work_dir = './work_dirs/retinanet_free_anchor_x101-32x4d_fpn_1x'
diff --git a/configs/gcnet/mask_rcnn_r16_gcb_c3-c5_r50_fpn_1x.py b/configs/gcnet/mask_rcnn_r16_gcb_c3-c5_r50_fpn_1x.py
deleted file mode 100644
index bce18b940a59b29f3d053f1c00e24db7f86e5bee..0000000000000000000000000000000000000000
--- a/configs/gcnet/mask_rcnn_r16_gcb_c3-c5_r50_fpn_1x.py
+++ /dev/null
@@ -1,192 +0,0 @@
-# model settings
-model = dict(
- type='MaskRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- gcb=dict(ratio=1. / 16., ),
- stage_with_gcb=(False, True, True, True)),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_r16_gcb_c3-c5_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/gcnet/mask_rcnn_r16_gcb_c3-c5_r50_fpn_syncbn_1x.py b/configs/gcnet/mask_rcnn_r16_gcb_c3-c5_r50_fpn_syncbn_1x.py
deleted file mode 100644
index acb15eb33a8c21b9d6eac549505f435ef69fe9e0..0000000000000000000000000000000000000000
--- a/configs/gcnet/mask_rcnn_r16_gcb_c3-c5_r50_fpn_syncbn_1x.py
+++ /dev/null
@@ -1,193 +0,0 @@
-# model settings
-model = dict(
- type='MaskRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='SyncBN', requires_grad=True),
- norm_eval=False,
- style='pytorch',
- gcb=dict(ratio=1. / 16., ),
- stage_with_gcb=(False, True, True, True)),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_r16_gcb_c3-c5_r50_fpn_syncbn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/gcnet/mask_rcnn_r4_gcb_c3-c5_r50_fpn_1x.py b/configs/gcnet/mask_rcnn_r4_gcb_c3-c5_r50_fpn_1x.py
deleted file mode 100644
index 41058e74a71eb510a4acc78773789f5f06d04427..0000000000000000000000000000000000000000
--- a/configs/gcnet/mask_rcnn_r4_gcb_c3-c5_r50_fpn_1x.py
+++ /dev/null
@@ -1,192 +0,0 @@
-# model settings
-model = dict(
- type='MaskRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- gcb=dict(ratio=1. / 4., ),
- stage_with_gcb=(False, True, True, True)),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_r4_gcb_c3-c5_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/gcnet/mask_rcnn_r4_gcb_c3-c5_r50_fpn_syncbn_1x.py b/configs/gcnet/mask_rcnn_r4_gcb_c3-c5_r50_fpn_syncbn_1x.py
deleted file mode 100644
index 9e047758cf37dfeaba98713077e33e2714379c65..0000000000000000000000000000000000000000
--- a/configs/gcnet/mask_rcnn_r4_gcb_c3-c5_r50_fpn_syncbn_1x.py
+++ /dev/null
@@ -1,193 +0,0 @@
-# model settings
-model = dict(
- type='MaskRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='SyncBN', requires_grad=True),
- norm_eval=False,
- style='pytorch',
- gcb=dict(ratio=1. / 4., ),
- stage_with_gcb=(False, True, True, True)),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_r4_gcb_c3-c5_r50_fpn_syncbn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/gcnet/mask_rcnn_r50_fpn_r16_gcb_c3-c5_1x_coco.py b/configs/gcnet/mask_rcnn_r50_fpn_r16_gcb_c3-c5_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..f84d93e939cc4ce742a3bdeffcd08dbb3c8bedcf
--- /dev/null
+++ b/configs/gcnet/mask_rcnn_r50_fpn_r16_gcb_c3-c5_1x_coco.py
@@ -0,0 +1,5 @@
+_base_ = '../mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ backbone=dict(
+ gcb=dict(ratio=1. / 16., ), stage_with_gcb=(False, True, True, True)))
+work_dir = './work_dirs/mask_rcnn_r16_gcb_c3-c5_r50_fpn_1x'
diff --git a/configs/gcnet/mask_rcnn_r50_fpn_r4_gcb_c3-c5_1x_coco.py b/configs/gcnet/mask_rcnn_r50_fpn_r4_gcb_c3-c5_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..d15ae123da5d545d80bdd4bbb5b7210516bc4011
--- /dev/null
+++ b/configs/gcnet/mask_rcnn_r50_fpn_r4_gcb_c3-c5_1x_coco.py
@@ -0,0 +1,5 @@
+_base_ = '../mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ backbone=dict(
+ gcb=dict(ratio=1. / 4., ), stage_with_gcb=(False, True, True, True)))
+work_dir = './work_dirs/mask_rcnn_r4_gcb_c3-c5_r50_fpn_1x'
diff --git a/configs/gcnet/mask_rcnn_r50_fpn_sbn_1x.py b/configs/gcnet/mask_rcnn_r50_fpn_sbn_1x.py
deleted file mode 100644
index f273a3a62c9f44f8b23a38ace06b278ac7f009fe..0000000000000000000000000000000000000000
--- a/configs/gcnet/mask_rcnn_r50_fpn_sbn_1x.py
+++ /dev/null
@@ -1,191 +0,0 @@
-# model settings
-model = dict(
- type='MaskRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='SyncBN', requires_grad=True),
- norm_eval=False,
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_r50_fpn_sbn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/gcnet/mask_rcnn_r50_fpn_syncbn-backbone_1x_coco.py b/configs/gcnet/mask_rcnn_r50_fpn_syncbn-backbone_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..a3f33c5254fd76445c728032830fdd40df9fedf8
--- /dev/null
+++ b/configs/gcnet/mask_rcnn_r50_fpn_syncbn-backbone_1x_coco.py
@@ -0,0 +1,5 @@
+_base_ = '../mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ backbone=dict(
+ norm_cfg=dict(type='SyncBN', requires_grad=True), norm_eval=False))
+work_dir = './work_dirs/mask_rcnn_r50_fpn_sbn_1x'
diff --git a/configs/gcnet/mask_rcnn_r50_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco.py b/configs/gcnet/mask_rcnn_r50_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..26560197170f9ed934cb16b1ab7c6314425e65e4
--- /dev/null
+++ b/configs/gcnet/mask_rcnn_r50_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco.py
@@ -0,0 +1,8 @@
+_base_ = '../mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ backbone=dict(
+ norm_cfg=dict(type='SyncBN', requires_grad=True),
+ norm_eval=False,
+ stage_with_gcb=(False, True, True, True),
+ gcb=dict(ratio=1. / 16., )))
+work_dir = './work_dirs/mask_rcnn_r16_gcb_c3-c5_r50_fpn_syncbn_1x'
diff --git a/configs/gcnet/mask_rcnn_r50_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco.py b/configs/gcnet/mask_rcnn_r50_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..eb0a9bb49c59c83a5fb136004b62b7ffd5efc7e1
--- /dev/null
+++ b/configs/gcnet/mask_rcnn_r50_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco.py
@@ -0,0 +1,8 @@
+_base_ = '../mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ backbone=dict(
+ norm_cfg=dict(type='SyncBN', requires_grad=True),
+ norm_eval=False,
+ gcb=dict(ratio=1. / 4., ),
+ stage_with_gcb=(False, True, True, True)))
+work_dir = './work_dirs/mask_rcnn_r4_gcb_c3-c5_r50_fpn_syncbn_1x'
diff --git a/configs/ghm/retinanet_ghm_r50_fpn_1x.py b/configs/ghm/retinanet_ghm_r50_fpn_1x.py
deleted file mode 100644
index 3126a430c7fdff1a818abe12b402b978acf02c16..0000000000000000000000000000000000000000
--- a/configs/ghm/retinanet_ghm_r50_fpn_1x.py
+++ /dev/null
@@ -1,132 +0,0 @@
-# model settings
-model = dict(
- type='RetinaNet',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5),
- bbox_head=dict(
- type='RetinaHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- octave_base_scale=4,
- scales_per_octave=3,
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[8, 16, 32, 64, 128],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='GHMC',
- bins=30,
- momentum=0.75,
- use_sigmoid=True,
- loss_weight=1.0),
- loss_bbox=dict(
- type='GHMR', mu=0.02, bins=10, momentum=0.7, loss_weight=10.0)))
-# training and testing settings
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/ghm'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/ghm/retinanet_ghm_r50_fpn_1x_coco.py b/configs/ghm/retinanet_ghm_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..42a921fdde064ccde81faa9bc8309e94acbf087a
--- /dev/null
+++ b/configs/ghm/retinanet_ghm_r50_fpn_1x_coco.py
@@ -0,0 +1,18 @@
+_base_ = '../retinanet/retinanet_r50_fpn_1x_coco.py'
+model = dict(
+ bbox_head=dict(
+ loss_cls=dict(
+ _delete_=True,
+ type='GHMC',
+ bins=30,
+ momentum=0.75,
+ use_sigmoid=True,
+ loss_weight=1.0),
+ loss_bbox=dict(
+ _delete_=True,
+ type='GHMR',
+ mu=0.02,
+ bins=10,
+ momentum=0.7,
+ loss_weight=10.0)))
+work_dir = './work_dirs/ghm'
diff --git a/configs/gn+ws/faster_rcnn_r50_fpn_gn_ws-all_1x_coco.py b/configs/gn+ws/faster_rcnn_r50_fpn_gn_ws-all_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..39d2480860429e2a7b7734fc41c2d2f8d750d361
--- /dev/null
+++ b/configs/gn+ws/faster_rcnn_r50_fpn_gn_ws-all_1x_coco.py
@@ -0,0 +1,13 @@
+_base_ = '../faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
+conv_cfg = dict(type='ConvWS')
+norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
+model = dict(
+ pretrained='open-mmlab://jhu/resnet50_gn_ws',
+ backbone=dict(conv_cfg=conv_cfg, norm_cfg=norm_cfg),
+ neck=dict(conv_cfg=conv_cfg, norm_cfg=norm_cfg),
+ bbox_head=dict(
+ type='Shared4Conv1FCBBoxHead',
+ conv_out_channels=256,
+ conv_cfg=conv_cfg,
+ norm_cfg=norm_cfg))
+work_dir = './work_dirs/faster_rcnn_r50_fpn_gn_ws_1x'
diff --git a/configs/gn+ws/faster_rcnn_r50_fpn_gn_ws_1x.py b/configs/gn+ws/faster_rcnn_r50_fpn_gn_ws_1x.py
deleted file mode 100644
index 2cd1969c56478e89e3268ecf7e6b2092f715354d..0000000000000000000000000000000000000000
--- a/configs/gn+ws/faster_rcnn_r50_fpn_gn_ws_1x.py
+++ /dev/null
@@ -1,182 +0,0 @@
-# model settings
-conv_cfg = dict(type='ConvWS')
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-model = dict(
- type='FasterRCNN',
- pretrained='open-mmlab://jhu/resnet50_gn_ws',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- style='pytorch',
- conv_cfg=conv_cfg,
- norm_cfg=norm_cfg),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5,
- conv_cfg=conv_cfg,
- norm_cfg=norm_cfg),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='ConvFCBBoxHead',
- num_shared_convs=4,
- num_shared_fcs=1,
- in_channels=256,
- conv_out_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- conv_cfg=conv_cfg,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_r50_fpn_gn_ws_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/gn+ws/mask_rcnn_r50_fpn_gn_ws-all_20_23_24e_coco.py b/configs/gn+ws/mask_rcnn_r50_fpn_gn_ws-all_20_23_24e_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..bb241407840152fdda64907d39bad790249c980a
--- /dev/null
+++ b/configs/gn+ws/mask_rcnn_r50_fpn_gn_ws-all_20_23_24e_coco.py
@@ -0,0 +1,5 @@
+_base_ = './mask_rcnn_r50_fpn_gn_ws-all_2x_coco.py'
+# learning policy
+lr_config = dict(step=[20, 23])
+total_epochs = 24
+work_dir = './work_dirs/mask_rcnn_r50_fpn_gn_ws_20_23_24e'
diff --git a/configs/gn+ws/mask_rcnn_r50_fpn_gn_ws-all_2x_coco.py b/configs/gn+ws/mask_rcnn_r50_fpn_gn_ws-all_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..da3f400941d3a298c12b0f2c5a4fb90225cc3461
--- /dev/null
+++ b/configs/gn+ws/mask_rcnn_r50_fpn_gn_ws-all_2x_coco.py
@@ -0,0 +1,17 @@
+_base_ = '../mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py'
+conv_cfg = dict(type='ConvWS')
+norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
+model = dict(
+ pretrained='open-mmlab://jhu/resnet50_gn_ws',
+ backbone=dict(conv_cfg=conv_cfg, norm_cfg=norm_cfg),
+ neck=dict(conv_cfg=conv_cfg, norm_cfg=norm_cfg),
+ bbox_head=dict(
+ type='Shared4Conv1FCBBoxHead',
+ conv_out_channels=256,
+ conv_cfg=conv_cfg,
+ norm_cfg=norm_cfg),
+ mask_head=dict(conv_cfg=conv_cfg, norm_cfg=norm_cfg))
+# learning policy
+lr_config = dict(step=[16, 22])
+total_epochs = 24
+work_dir = './work_dirs/mask_rcnn_r50_fpn_gn_ws_2x'
diff --git a/configs/gn+ws/mask_rcnn_r50_fpn_gn_ws_20_23_24e.py b/configs/gn+ws/mask_rcnn_r50_fpn_gn_ws_20_23_24e.py
deleted file mode 100644
index 5a698f269d767b47b49cfb023357b933d52f3bb1..0000000000000000000000000000000000000000
--- a/configs/gn+ws/mask_rcnn_r50_fpn_gn_ws_20_23_24e.py
+++ /dev/null
@@ -1,201 +0,0 @@
-# model settings
-conv_cfg = dict(type='ConvWS')
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-model = dict(
- type='MaskRCNN',
- pretrained='open-mmlab://jhu/resnet50_gn_ws',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- style='pytorch',
- conv_cfg=conv_cfg,
- norm_cfg=norm_cfg),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5,
- conv_cfg=conv_cfg,
- norm_cfg=norm_cfg),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='ConvFCBBoxHead',
- num_shared_convs=4,
- num_shared_fcs=1,
- in_channels=256,
- conv_out_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- conv_cfg=conv_cfg,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- conv_cfg=conv_cfg,
- norm_cfg=norm_cfg,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[20, 23])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_r50_fpn_gn_ws_20_23_24e'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/gn+ws/mask_rcnn_r50_fpn_gn_ws_2x.py b/configs/gn+ws/mask_rcnn_r50_fpn_gn_ws_2x.py
deleted file mode 100644
index 00dc587a33b829808d241415731511649002779e..0000000000000000000000000000000000000000
--- a/configs/gn+ws/mask_rcnn_r50_fpn_gn_ws_2x.py
+++ /dev/null
@@ -1,201 +0,0 @@
-# model settings
-conv_cfg = dict(type='ConvWS')
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-model = dict(
- type='MaskRCNN',
- pretrained='open-mmlab://jhu/resnet50_gn_ws',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- style='pytorch',
- conv_cfg=conv_cfg,
- norm_cfg=norm_cfg),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5,
- conv_cfg=conv_cfg,
- norm_cfg=norm_cfg),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='ConvFCBBoxHead',
- num_shared_convs=4,
- num_shared_fcs=1,
- in_channels=256,
- conv_out_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- conv_cfg=conv_cfg,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- conv_cfg=conv_cfg,
- norm_cfg=norm_cfg,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_r50_fpn_gn_ws_2x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/gn+ws/mask_rcnn_x101_32x4d_fpn_gn_ws-all_2x_coco.py b/configs/gn+ws/mask_rcnn_x101_32x4d_fpn_gn_ws-all_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..40c067f2db725ecadfb550b614a4f311b8ac34e9
--- /dev/null
+++ b/configs/gn+ws/mask_rcnn_x101_32x4d_fpn_gn_ws-all_2x_coco.py
@@ -0,0 +1,18 @@
+_base_ = './mask_rcnn_r50_fpn_gn_ws-all_2x_coco.py'
+# model settings
+conv_cfg = dict(type='ConvWS')
+norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
+model = dict(
+ pretrained='open-mmlab://jhu/resnext101_32x4d_gn_ws',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=32,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ style='pytorch',
+ conv_cfg=conv_cfg,
+ norm_cfg=norm_cfg))
+work_dir = './work_dirs/mask_rcnn_x101_32x4d_fpn_gn_ws_2x'
diff --git a/configs/gn+ws/mask_rcnn_x101_32x4d_fpn_gn_ws_2x.py b/configs/gn+ws/mask_rcnn_x101_32x4d_fpn_gn_ws_2x.py
deleted file mode 100644
index 11077506df6909bf914eb8eb84c3d512cc6a83ec..0000000000000000000000000000000000000000
--- a/configs/gn+ws/mask_rcnn_x101_32x4d_fpn_gn_ws_2x.py
+++ /dev/null
@@ -1,203 +0,0 @@
-# model settings
-conv_cfg = dict(type='ConvWS')
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-model = dict(
- type='MaskRCNN',
- pretrained='open-mmlab://jhu/resnext101_32x4d_gn_ws',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=32,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- style='pytorch',
- conv_cfg=conv_cfg,
- norm_cfg=norm_cfg),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5,
- conv_cfg=conv_cfg,
- norm_cfg=norm_cfg),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='ConvFCBBoxHead',
- num_shared_convs=4,
- num_shared_fcs=1,
- in_channels=256,
- conv_out_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- conv_cfg=conv_cfg,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- conv_cfg=conv_cfg,
- norm_cfg=norm_cfg,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_x101_32x4d_fpn_gn_ws_2x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/gn/mask_rcnn_r101_fpn_gn-all_2x_coco.py b/configs/gn/mask_rcnn_r101_fpn_gn-all_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..0ea44d2eb09675b4dc47ffa6fde370382cbc8e88
--- /dev/null
+++ b/configs/gn/mask_rcnn_r101_fpn_gn-all_2x_coco.py
@@ -0,0 +1,4 @@
+_base_ = './mask_rcnn_r50_fpn_gn-all_2x_coco.py'
+model = dict(
+ pretrained='open-mmlab://detectron/resnet101_gn', backbone=dict(depth=101))
+work_dir = './work_dirs/mask_rcnn_r101_fpn_gn_2x'
diff --git a/configs/gn/mask_rcnn_r101_fpn_gn_2x.py b/configs/gn/mask_rcnn_r101_fpn_gn_2x.py
deleted file mode 100644
index d13670960ee48dd44eaa45102aa5bfd77e355fb8..0000000000000000000000000000000000000000
--- a/configs/gn/mask_rcnn_r101_fpn_gn_2x.py
+++ /dev/null
@@ -1,197 +0,0 @@
-# model settings
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-
-model = dict(
- type='MaskRCNN',
- pretrained='open-mmlab://detectron/resnet101_gn',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- style='pytorch',
- norm_cfg=norm_cfg),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5,
- norm_cfg=norm_cfg),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='ConvFCBBoxHead',
- num_shared_convs=4,
- num_shared_fcs=1,
- in_channels=256,
- conv_out_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- norm_cfg=norm_cfg,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_r101_fpn_gn_2x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/gn/mask_rcnn_r50_fpn_gn-all_2x_coco.py b/configs/gn/mask_rcnn_r50_fpn_gn-all_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..6fd8b4c90066addb2d7bd3d5565bf953ca68a091
--- /dev/null
+++ b/configs/gn/mask_rcnn_r50_fpn_gn-all_2x_coco.py
@@ -0,0 +1,46 @@
+_base_ = '../mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py'
+norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
+model = dict(
+ pretrained='open-mmlab://detectron/resnet50_gn',
+ backbone=dict(norm_cfg=norm_cfg),
+ neck=dict(norm_cfg=norm_cfg),
+ bbox_head=dict(
+ type='Shared4Conv1FCBBoxHead',
+ conv_out_channels=256,
+ norm_cfg=norm_cfg),
+ mask_head=dict(norm_cfg=norm_cfg))
+img_norm_cfg = dict(
+ mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(1333, 800),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ train=dict(pipeline=train_pipeline),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+# learning policy
+lr_config = dict(step=[16, 22])
+total_epochs = 24
+work_dir = './work_dirs/mask_rcnn_r50_fpn_gn_2x'
diff --git a/configs/gn/mask_rcnn_r50_fpn_gn-all_contrib_2x_coco.py b/configs/gn/mask_rcnn_r50_fpn_gn-all_contrib_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..e01585675f003716ae0df59cff958532bdc20926
--- /dev/null
+++ b/configs/gn/mask_rcnn_r50_fpn_gn-all_contrib_2x_coco.py
@@ -0,0 +1,15 @@
+_base_ = '../mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py'
+norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
+model = dict(
+ pretrained='open-mmlab://contrib/resnet50_gn',
+ backbone=dict(norm_cfg=norm_cfg),
+ neck=dict(norm_cfg=norm_cfg),
+ bbox_head=dict(
+ type='Shared4Conv1FCBBoxHead',
+ conv_out_channels=256,
+ norm_cfg=norm_cfg),
+ mask_head=dict(norm_cfg=norm_cfg))
+# learning policy
+lr_config = dict(step=[16, 22])
+total_epochs = 24
+work_dir = './work_dirs/mask_rcnn_r50_fpn_gn_contrib_2x'
diff --git a/configs/gn/mask_rcnn_r50_fpn_gn_2x.py b/configs/gn/mask_rcnn_r50_fpn_gn_2x.py
deleted file mode 100644
index e0e09b5687b884200f1f9315c737c9b124505e81..0000000000000000000000000000000000000000
--- a/configs/gn/mask_rcnn_r50_fpn_gn_2x.py
+++ /dev/null
@@ -1,197 +0,0 @@
-# model settings
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-
-model = dict(
- type='MaskRCNN',
- pretrained='open-mmlab://detectron/resnet50_gn',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- style='pytorch',
- norm_cfg=norm_cfg),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5,
- norm_cfg=norm_cfg),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='ConvFCBBoxHead',
- num_shared_convs=4,
- num_shared_fcs=1,
- in_channels=256,
- conv_out_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- norm_cfg=norm_cfg,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_r50_fpn_gn_2x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/gn/mask_rcnn_r50_fpn_gn_contrib_2x.py b/configs/gn/mask_rcnn_r50_fpn_gn_contrib_2x.py
deleted file mode 100644
index 271f6d25d12c43329df74dc02cd5b38dee560058..0000000000000000000000000000000000000000
--- a/configs/gn/mask_rcnn_r50_fpn_gn_contrib_2x.py
+++ /dev/null
@@ -1,197 +0,0 @@
-# model settings
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-
-model = dict(
- type='MaskRCNN',
- pretrained='open-mmlab://contrib/resnet50_gn',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- style='pytorch',
- norm_cfg=norm_cfg),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5,
- norm_cfg=norm_cfg),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='ConvFCBBoxHead',
- num_shared_convs=4,
- num_shared_fcs=1,
- in_channels=256,
- conv_out_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- norm_cfg=norm_cfg,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_r50_fpn_gn_contrib_2x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/grid_rcnn/grid_rcnn_gn_head_x101_32x4d_fpn_2x.py b/configs/grid_rcnn/grid_rcnn_gn_head_x101_32x4d_fpn_2x.py
deleted file mode 100644
index fd459b0cb3832036e700fd93479e7b136058a389..0000000000000000000000000000000000000000
--- a/configs/grid_rcnn/grid_rcnn_gn_head_x101_32x4d_fpn_2x.py
+++ /dev/null
@@ -1,189 +0,0 @@
-# model settings
-model = dict(
- type='GridRCNN',
- pretrained='open-mmlab://resnext101_32x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=32,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- with_reg=False,
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False),
- grid_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- grid_head=dict(
- type='GridHead',
- grid_points=9,
- num_convs=8,
- in_channels=256,
- point_feat_channels=64,
- norm_cfg=dict(type='GN', num_groups=36),
- loss_grid=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=15)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_radius=1,
- pos_weight=-1,
- max_num_grid=192,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.03, nms=dict(type='nms', iou_thr=0.3), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=None)
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=3665,
- warmup_ratio=1.0 / 80,
- step=[17, 23])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 25
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/grid_rcnn_gn_head_x101_32x4d_fpn_2x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/grid_rcnn/grid_rcnn_gn_head_r50_fpn_2x.py b/configs/grid_rcnn/grid_rcnn_r50_fpn_gn-head_2x_coco.py
similarity index 63%
rename from configs/grid_rcnn/grid_rcnn_gn_head_r50_fpn_2x.py
rename to configs/grid_rcnn/grid_rcnn_r50_fpn_gn-head_2x_coco.py
index 810e6ddecd6ede880b4ee2cc92284799d921d3d3..e835ab7e805b7981441b079d5a4616c96467af42 100644
--- a/configs/grid_rcnn/grid_rcnn_gn_head_r50_fpn_2x.py
+++ b/configs/grid_rcnn/grid_rcnn_r50_fpn_gn-head_2x_coco.py
@@ -1,3 +1,6 @@
+_base_ = [
+ '../_base_/datasets/coco_detection.py', '../_base_/default_runtime.py'
+]
# model settings
model = dict(
type='GridRCNN',
@@ -33,9 +36,8 @@ model = dict(
out_channels=256,
featmap_strides=[4, 8, 16, 32]),
bbox_head=dict(
- type='SharedFCBBoxHead',
+ type='Shared2FCBBoxHead',
with_reg=False,
- num_fcs=2,
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
@@ -109,55 +111,6 @@ test_cfg = dict(
min_bbox_size=0),
rcnn=dict(
score_thr=0.03, nms=dict(type='nms', iou_thr=0.3), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
# optimizer
optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
optimizer_config = dict(grad_clip=None)
@@ -168,20 +121,5 @@ lr_config = dict(
warmup_iters=3665,
warmup_ratio=1.0 / 80,
step=[17, 23])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
total_epochs = 25
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
work_dir = './work_dirs/grid_rcnn_gn_head_r50_fpn_2x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/grid_rcnn/grid_rcnn_x101_32x4d_fpn_gn-head_2x_coco.py b/configs/grid_rcnn/grid_rcnn_x101_32x4d_fpn_gn-head_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..c42aed798605f9b8a1db8e06cfbf320df96d8622
--- /dev/null
+++ b/configs/grid_rcnn/grid_rcnn_x101_32x4d_fpn_gn-head_2x_coco.py
@@ -0,0 +1,24 @@
+_base_ = './grid_rcnn_r50_fpn_gn-head_2x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_32x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=32,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ style='pytorch'))
+# optimizer
+optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
+optimizer_config = dict(grad_clip=None)
+# learning policy
+lr_config = dict(
+ policy='step',
+ warmup='linear',
+ warmup_iters=3665,
+ warmup_ratio=1.0 / 80,
+ step=[17, 23])
+total_epochs = 25
+work_dir = './work_dirs/grid_rcnn_gn_head_x101_32x4d_fpn_2x'
diff --git a/configs/guided_anchoring/ga_fast_r50_caffe_fpn_1x.py b/configs/guided_anchoring/ga_fast_r50_caffe_fpn_1x.py
deleted file mode 100644
index d5a4753ebec622e116eac0c5ba44dd081348b9da..0000000000000000000000000000000000000000
--- a/configs/guided_anchoring/ga_fast_r50_caffe_fpn_1x.py
+++ /dev/null
@@ -1,137 +0,0 @@
-# model settings
-model = dict(
- type='FastRCNN',
- pretrained='open-mmlab://resnet50_caffe',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=False),
- norm_eval=True,
- style='caffe'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rcnn=dict(
- score_thr=1e-3, nms=dict(type='nms', iou_thr=0.5), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadProposals', num_max_proposals=300),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'proposals', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadProposals', num_max_proposals=None),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img', 'proposals']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- proposal_file=data_root + 'proposals/ga_rpn_r50_fpn_1x_train2017.pkl',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- proposal_file=data_root + 'proposals/ga_rpn_r50_fpn_1x_val2017.pkl',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- proposal_file=data_root + 'proposals/ga_rpn_r50_fpn_1x_val2017.pkl',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/ga_fast_rcnn_r50_caffe_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/guided_anchoring/ga_fast_r50_caffe_fpn_1x_coco.py b/configs/guided_anchoring/ga_fast_r50_caffe_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..3056f2247ad7c313065c1769a907cdc09f7f4c12
--- /dev/null
+++ b/configs/guided_anchoring/ga_fast_r50_caffe_fpn_1x_coco.py
@@ -0,0 +1,60 @@
+_base_ = '../fast_rcnn/fast_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnet50_caffe',
+ backbone=dict(
+ type='ResNet',
+ depth=50,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=False),
+ style='caffe'),
+ bbox_head=dict(target_stds=[0.05, 0.05, 0.1, 0.1]))
+# model training and testing settings
+train_cfg = dict(
+ rcnn=dict(
+ assigner=dict(pos_iou_thr=0.6, neg_iou_thr=0.6, min_pos_iou=0.6),
+ sampler=dict(num=256)))
+test_cfg = dict(rcnn=dict(score_thr=1e-3))
+dataset_type = 'CocoDataset'
+data_root = 'data/coco/'
+img_norm_cfg = dict(
+ mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadProposals', num_max_proposals=300),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'proposals', 'gt_bboxes', 'gt_labels']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadProposals', num_max_proposals=None),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(1333, 800),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img', 'proposals']),
+ ])
+]
+data = dict(
+ train=dict(
+ proposal_file=data_root + 'proposals/ga_rpn_r50_fpn_1x_train2017.pkl',
+ pipeline=train_pipeline),
+ val=dict(
+ proposal_file=data_root + 'proposals/ga_rpn_r50_fpn_1x_val2017.pkl',
+ pipeline=test_pipeline),
+ test=dict(
+ proposal_file=data_root + 'proposals/ga_rpn_r50_fpn_1x_val2017.pkl',
+ pipeline=test_pipeline))
+work_dir = './work_dirs/ga_fast_rcnn_r50_caffe_fpn_1x'
diff --git a/configs/guided_anchoring/ga_faster_r50_caffe_fpn_1x.py b/configs/guided_anchoring/ga_faster_r50_caffe_fpn_1x.py
deleted file mode 100644
index 57da8d486b43f4e588587bbacf3b1652c0e843e0..0000000000000000000000000000000000000000
--- a/configs/guided_anchoring/ga_faster_r50_caffe_fpn_1x.py
+++ /dev/null
@@ -1,200 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='open-mmlab://resnet50_caffe',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=False),
- norm_eval=True,
- style='caffe'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='GARPNHead',
- in_channels=256,
- feat_channels=256,
- octave_base_scale=8,
- scales_per_octave=3,
- octave_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- anchor_base_sizes=None,
- anchoring_means=[.0, .0, .0, .0],
- anchoring_stds=[0.07, 0.07, 0.14, 0.14],
- target_means=(.0, .0, .0, .0),
- target_stds=[0.07, 0.07, 0.11, 0.11],
- loc_filter_thr=0.01,
- loss_loc=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_shape=dict(type='BoundedIoULoss', beta=0.2, loss_weight=1.0),
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- ga_assigner=dict(
- type='ApproxMaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- ga_sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=-1,
- pos_weight=-1,
- center_ratio=0.2,
- ignore_ratio=0.5,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=300,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=300,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=1e-3, nms=dict(type='nms', iou_thr=0.5), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/ga_faster_rcnn_r50_caffe_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/guided_anchoring/ga_faster_r50_caffe_fpn_1x_coco.py b/configs/guided_anchoring/ga_faster_r50_caffe_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..0fd06b184c2e66d17793b9b0ff287f0a45f08949
--- /dev/null
+++ b/configs/guided_anchoring/ga_faster_r50_caffe_fpn_1x_coco.py
@@ -0,0 +1,52 @@
+_base_ = '../faster_rcnn/faster_rcnn_r50_caffe_fpn_1x_coco.py'
+model = dict(
+ rpn_head=dict(
+ _delete_=True,
+ type='GARPNHead',
+ in_channels=256,
+ feat_channels=256,
+ octave_base_scale=8,
+ scales_per_octave=3,
+ octave_ratios=[0.5, 1.0, 2.0],
+ anchor_strides=[4, 8, 16, 32, 64],
+ anchor_base_sizes=None,
+ anchoring_means=[.0, .0, .0, .0],
+ anchoring_stds=[0.07, 0.07, 0.14, 0.14],
+ target_means=(.0, .0, .0, .0),
+ target_stds=[0.07, 0.07, 0.11, 0.11],
+ loc_filter_thr=0.01,
+ loss_loc=dict(
+ type='FocalLoss',
+ use_sigmoid=True,
+ gamma=2.0,
+ alpha=0.25,
+ loss_weight=1.0),
+ loss_shape=dict(type='BoundedIoULoss', beta=0.2, loss_weight=1.0),
+ loss_cls=dict(
+ type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
+ loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
+ bbox_head=dict(target_stds=[0.05, 0.05, 0.1, 0.1]))
+# model training and testing settings
+train_cfg = dict(
+ rpn=dict(
+ ga_assigner=dict(
+ type='ApproxMaxIoUAssigner',
+ pos_iou_thr=0.7,
+ neg_iou_thr=0.3,
+ min_pos_iou=0.3,
+ ignore_iof_thr=-1),
+ ga_sampler=dict(
+ type='RandomSampler',
+ num=256,
+ pos_fraction=0.5,
+ neg_pos_ub=-1,
+ add_gt_as_proposals=False),
+ allowed_border=-1,
+ center_ratio=0.2,
+ ignore_ratio=0.5),
+ rpn_proposal=dict(max_num=300),
+ rcnn=dict(
+ assigner=dict(pos_iou_thr=0.6, neg_iou_thr=0.6, min_pos_iou=0.6),
+ sampler=dict(type='RandomSampler', num=256)))
+test_cfg = dict(rpn=dict(max_num=300), rcnn=dict(score_thr=1e-3))
+work_dir = './work_dirs/ga_faster_rcnn_r50_caffe_fpn_1x'
diff --git a/configs/guided_anchoring/ga_faster_r50_fpn_1x_coco.py b/configs/guided_anchoring/ga_faster_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..482845e8e57ff325bb447797f1ddb0e381934299
--- /dev/null
+++ b/configs/guided_anchoring/ga_faster_r50_fpn_1x_coco.py
@@ -0,0 +1,52 @@
+_base_ = '../faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ rpn_head=dict(
+ _delete_=True,
+ type='GARPNHead',
+ in_channels=256,
+ feat_channels=256,
+ octave_base_scale=8,
+ scales_per_octave=3,
+ octave_ratios=[0.5, 1.0, 2.0],
+ anchor_strides=[4, 8, 16, 32, 64],
+ anchor_base_sizes=None,
+ anchoring_means=[.0, .0, .0, .0],
+ anchoring_stds=[0.07, 0.07, 0.14, 0.14],
+ target_means=(.0, .0, .0, .0),
+ target_stds=[0.07, 0.07, 0.11, 0.11],
+ loc_filter_thr=0.01,
+ loss_loc=dict(
+ type='FocalLoss',
+ use_sigmoid=True,
+ gamma=2.0,
+ alpha=0.25,
+ loss_weight=1.0),
+ loss_shape=dict(type='BoundedIoULoss', beta=0.2, loss_weight=1.0),
+ loss_cls=dict(
+ type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
+ loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
+ bbox_head=dict(target_stds=[0.05, 0.05, 0.1, 0.1]))
+# model training and testing settings
+train_cfg = dict(
+ rpn=dict(
+ ga_assigner=dict(
+ type='ApproxMaxIoUAssigner',
+ pos_iou_thr=0.7,
+ neg_iou_thr=0.3,
+ min_pos_iou=0.3,
+ ignore_iof_thr=-1),
+ ga_sampler=dict(
+ type='RandomSampler',
+ num=256,
+ pos_fraction=0.5,
+ neg_pos_ub=-1,
+ add_gt_as_proposals=False),
+ allowed_border=-1,
+ center_ratio=0.2,
+ ignore_ratio=0.5),
+ rpn_proposal=dict(max_num=300),
+ rcnn=dict(
+ assigner=dict(pos_iou_thr=0.6, neg_iou_thr=0.6, min_pos_iou=0.6),
+ sampler=dict(type='RandomSampler', num=256)))
+test_cfg = dict(rpn=dict(max_num=300), rcnn=dict(score_thr=1e-3))
+work_dir = './work_dirs/ga_faster_rcnn_r50_fpn_1x'
diff --git a/configs/guided_anchoring/ga_faster_x101_32x4d_fpn_1x.py b/configs/guided_anchoring/ga_faster_x101_32x4d_fpn_1x.py
deleted file mode 100644
index 57a39228b5ce84c91367e102826875eb8fec4f44..0000000000000000000000000000000000000000
--- a/configs/guided_anchoring/ga_faster_x101_32x4d_fpn_1x.py
+++ /dev/null
@@ -1,201 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='open-mmlab://resnext101_32x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=32,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='GARPNHead',
- in_channels=256,
- feat_channels=256,
- octave_base_scale=8,
- scales_per_octave=3,
- octave_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- anchor_base_sizes=None,
- anchoring_means=[.0, .0, .0, .0],
- anchoring_stds=[0.07, 0.07, 0.14, 0.14],
- target_means=(.0, .0, .0, .0),
- target_stds=[0.07, 0.07, 0.11, 0.11],
- loc_filter_thr=0.01,
- loss_loc=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_shape=dict(type='BoundedIoULoss', beta=0.2, loss_weight=1.0),
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- ga_assigner=dict(
- type='ApproxMaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- ga_sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=-1,
- pos_weight=-1,
- center_ratio=0.2,
- ignore_ratio=0.5,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=300,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=300,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=1e-3, nms=dict(type='nms', iou_thr=0.5), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/ga_faster_rcnn_x101_32x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/guided_anchoring/ga_faster_x101_32x4d_fpn_1x_coco.py b/configs/guided_anchoring/ga_faster_x101_32x4d_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..f0a55e39702c510ea6a654726340108ff9335f67
--- /dev/null
+++ b/configs/guided_anchoring/ga_faster_x101_32x4d_fpn_1x_coco.py
@@ -0,0 +1,14 @@
+_base_ = './ga_faster_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_32x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=32,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+work_dir = './work_dirs/ga_faster_rcnn_x101_32x4d_fpn_1x'
diff --git a/configs/guided_anchoring/ga_retinanet_r50_caffe_fpn_1x.py b/configs/guided_anchoring/ga_retinanet_r50_caffe_fpn_1x.py
deleted file mode 100644
index 084b851a87303be5aef934d603f6e287dbccbff7..0000000000000000000000000000000000000000
--- a/configs/guided_anchoring/ga_retinanet_r50_caffe_fpn_1x.py
+++ /dev/null
@@ -1,157 +0,0 @@
-# model settings
-model = dict(
- type='RetinaNet',
- pretrained='open-mmlab://resnet50_caffe',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=False),
- norm_eval=True,
- style='caffe'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5),
- bbox_head=dict(
- type='GARetinaHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- octave_base_scale=4,
- scales_per_octave=3,
- octave_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[8, 16, 32, 64, 128],
- anchor_base_sizes=None,
- anchoring_means=[.0, .0, .0, .0],
- anchoring_stds=[1.0, 1.0, 1.0, 1.0],
- target_means=(.0, .0, .0, .0),
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loc_filter_thr=0.01,
- loss_loc=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_shape=dict(type='BoundedIoULoss', beta=0.2, loss_weight=1.0),
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=0.04, loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict(
- ga_assigner=dict(
- type='ApproxMaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0.4,
- ignore_iof_thr=-1),
- ga_sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- center_ratio=0.2,
- ignore_ratio=0.5,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/ga_retinanet_r50_caffe_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/guided_anchoring/ga_retinanet_r50_caffe_fpn_1x_coco.py b/configs/guided_anchoring/ga_retinanet_r50_caffe_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..63da4ec1de981fcc452c7f3c78aa09a9b17833e4
--- /dev/null
+++ b/configs/guided_anchoring/ga_retinanet_r50_caffe_fpn_1x_coco.py
@@ -0,0 +1,51 @@
+_base_ = '../retinanet/retinanet_r50_caffe_fpn_1x_coco.py'
+model = dict(
+ bbox_head=dict(
+ _delete_=True,
+ type='GARetinaHead',
+ num_classes=81,
+ in_channels=256,
+ stacked_convs=4,
+ feat_channels=256,
+ octave_base_scale=4,
+ scales_per_octave=3,
+ octave_ratios=[0.5, 1.0, 2.0],
+ anchor_strides=[8, 16, 32, 64, 128],
+ anchor_base_sizes=None,
+ anchoring_means=[.0, .0, .0, .0],
+ anchoring_stds=[1.0, 1.0, 1.0, 1.0],
+ target_means=(.0, .0, .0, .0),
+ target_stds=[1.0, 1.0, 1.0, 1.0],
+ loc_filter_thr=0.01,
+ loss_loc=dict(
+ type='FocalLoss',
+ use_sigmoid=True,
+ gamma=2.0,
+ alpha=0.25,
+ loss_weight=1.0),
+ loss_shape=dict(type='BoundedIoULoss', beta=0.2, loss_weight=1.0),
+ loss_cls=dict(
+ type='FocalLoss',
+ use_sigmoid=True,
+ gamma=2.0,
+ alpha=0.25,
+ loss_weight=1.0),
+ loss_bbox=dict(type='SmoothL1Loss', beta=0.04, loss_weight=1.0)))
+# training and testing settings
+train_cfg = dict(
+ ga_assigner=dict(
+ type='ApproxMaxIoUAssigner',
+ pos_iou_thr=0.5,
+ neg_iou_thr=0.4,
+ min_pos_iou=0.4,
+ ignore_iof_thr=-1),
+ ga_sampler=dict(
+ type='RandomSampler',
+ num=256,
+ pos_fraction=0.5,
+ neg_pos_ub=-1,
+ add_gt_as_proposals=False),
+ assigner=dict(neg_iou_thr=0.5, min_pos_iou=0.0),
+ center_ratio=0.2,
+ ignore_ratio=0.5)
+work_dir = './work_dirs/ga_retinanet_r50_caffe_fpn_1x'
diff --git a/configs/guided_anchoring/ga_retinanet_r50_fpn_1x_coco.py b/configs/guided_anchoring/ga_retinanet_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..35f1d33e299b1c4faa1aa7612163ecf5a614c005
--- /dev/null
+++ b/configs/guided_anchoring/ga_retinanet_r50_fpn_1x_coco.py
@@ -0,0 +1,51 @@
+_base_ = '../retinanet/retinanet_r50_fpn_1x_coco.py'
+model = dict(
+ bbox_head=dict(
+ _delete_=True,
+ type='GARetinaHead',
+ num_classes=81,
+ in_channels=256,
+ stacked_convs=4,
+ feat_channels=256,
+ octave_base_scale=4,
+ scales_per_octave=3,
+ octave_ratios=[0.5, 1.0, 2.0],
+ anchor_strides=[8, 16, 32, 64, 128],
+ anchor_base_sizes=None,
+ anchoring_means=[.0, .0, .0, .0],
+ anchoring_stds=[1.0, 1.0, 1.0, 1.0],
+ target_means=(.0, .0, .0, .0),
+ target_stds=[1.0, 1.0, 1.0, 1.0],
+ loc_filter_thr=0.01,
+ loss_loc=dict(
+ type='FocalLoss',
+ use_sigmoid=True,
+ gamma=2.0,
+ alpha=0.25,
+ loss_weight=1.0),
+ loss_shape=dict(type='BoundedIoULoss', beta=0.2, loss_weight=1.0),
+ loss_cls=dict(
+ type='FocalLoss',
+ use_sigmoid=True,
+ gamma=2.0,
+ alpha=0.25,
+ loss_weight=1.0),
+ loss_bbox=dict(type='SmoothL1Loss', beta=0.04, loss_weight=1.0)))
+# training and testing settings
+train_cfg = dict(
+ ga_assigner=dict(
+ type='ApproxMaxIoUAssigner',
+ pos_iou_thr=0.5,
+ neg_iou_thr=0.4,
+ min_pos_iou=0.4,
+ ignore_iof_thr=-1),
+ ga_sampler=dict(
+ type='RandomSampler',
+ num=256,
+ pos_fraction=0.5,
+ neg_pos_ub=-1,
+ add_gt_as_proposals=False),
+ assigner=dict(neg_iou_thr=0.5, min_pos_iou=0.0),
+ center_ratio=0.2,
+ ignore_ratio=0.5)
+work_dir = './work_dirs/ga_retinanet_r50_fpn_1x'
diff --git a/configs/guided_anchoring/ga_retinanet_x101_32x4d_fpn_1x.py b/configs/guided_anchoring/ga_retinanet_x101_32x4d_fpn_1x.py
deleted file mode 100644
index a47fad03aa128d24ad07522cd9a88f5698df3893..0000000000000000000000000000000000000000
--- a/configs/guided_anchoring/ga_retinanet_x101_32x4d_fpn_1x.py
+++ /dev/null
@@ -1,158 +0,0 @@
-# model settings
-model = dict(
- type='RetinaNet',
- pretrained='open-mmlab://resnext101_32x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=32,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5),
- bbox_head=dict(
- type='GARetinaHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- octave_base_scale=4,
- scales_per_octave=3,
- octave_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[8, 16, 32, 64, 128],
- anchor_base_sizes=None,
- anchoring_means=[.0, .0, .0, .0],
- anchoring_stds=[1.0, 1.0, 1.0, 1.0],
- target_means=(.0, .0, .0, .0),
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loc_filter_thr=0.01,
- loss_loc=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_shape=dict(type='BoundedIoULoss', beta=0.2, loss_weight=1.0),
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=0.04, loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict(
- ga_assigner=dict(
- type='ApproxMaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0.4,
- ignore_iof_thr=-1),
- ga_sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- center_ratio=0.2,
- ignore_ratio=0.5,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/ga_retinanet_x101_32x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/guided_anchoring/ga_retinanet_x101_32x4d_fpn_1x_coco.py b/configs/guided_anchoring/ga_retinanet_x101_32x4d_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..8889358e6bd77432529252ea9d1c8ccc3f0584cc
--- /dev/null
+++ b/configs/guided_anchoring/ga_retinanet_x101_32x4d_fpn_1x_coco.py
@@ -0,0 +1,14 @@
+_base_ = './ga_retinanet_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_32x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=32,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+work_dir = './work_dirs/ga_retinanet_x101_32x4d_fpn_1x'
diff --git a/configs/guided_anchoring/ga_rpn_r101_caffe_rpn_1x.py b/configs/guided_anchoring/ga_rpn_r101_caffe_rpn_1x.py
deleted file mode 100644
index 6629f12bb8d24916e89be2923aef6d5abf3a6497..0000000000000000000000000000000000000000
--- a/configs/guided_anchoring/ga_rpn_r101_caffe_rpn_1x.py
+++ /dev/null
@@ -1,158 +0,0 @@
-# model settings
-model = dict(
- type='RPN',
- pretrained='open-mmlab://resnet101_caffe',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=False),
- norm_eval=True,
- style='caffe'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='GARPNHead',
- in_channels=256,
- feat_channels=256,
- octave_base_scale=8,
- scales_per_octave=3,
- octave_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- anchor_base_sizes=None,
- anchoring_means=[.0, .0, .0, .0],
- anchoring_stds=[0.07, 0.07, 0.14, 0.14],
- target_means=(.0, .0, .0, .0),
- target_stds=[0.07, 0.07, 0.11, 0.11],
- loc_filter_thr=0.01,
- loss_loc=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_shape=dict(type='BoundedIoULoss', beta=0.2, loss_weight=1.0),
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- ga_assigner=dict(
- type='ApproxMaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- ga_sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=-1,
- pos_weight=-1,
- center_ratio=0.2,
- ignore_ratio=0.5,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_label=False),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='proposal_fast')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-# runner configs
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/ga_rpn_r101_caffe_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/guided_anchoring/ga_rpn_r101_caffe_rpn_1x_coco.py b/configs/guided_anchoring/ga_rpn_r101_caffe_rpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..69b91b53b941d69cc575027ed11296fa74cf8950
--- /dev/null
+++ b/configs/guided_anchoring/ga_rpn_r101_caffe_rpn_1x_coco.py
@@ -0,0 +1,5 @@
+_base_ = './ga_rpn_r50_caffe_fpn_1x_coco.py'
+# model settings
+model = dict(
+ pretrained='open-mmlab://resnet101_caffe', backbone=dict(depth=101))
+work_dir = './work_dirs/ga_rpn_r101_caffe_fpn_1x'
diff --git a/configs/guided_anchoring/ga_rpn_r50_caffe_fpn_1x.py b/configs/guided_anchoring/ga_rpn_r50_caffe_fpn_1x.py
deleted file mode 100644
index 507849e3c59a47cf089f5af957d1093902da8a7c..0000000000000000000000000000000000000000
--- a/configs/guided_anchoring/ga_rpn_r50_caffe_fpn_1x.py
+++ /dev/null
@@ -1,158 +0,0 @@
-# model settings
-model = dict(
- type='RPN',
- pretrained='open-mmlab://resnet50_caffe',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=False),
- norm_eval=True,
- style='caffe'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='GARPNHead',
- in_channels=256,
- feat_channels=256,
- octave_base_scale=8,
- scales_per_octave=3,
- octave_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- anchor_base_sizes=None,
- anchoring_means=[.0, .0, .0, .0],
- anchoring_stds=[0.07, 0.07, 0.14, 0.14],
- target_means=(.0, .0, .0, .0),
- target_stds=[0.07, 0.07, 0.11, 0.11],
- loc_filter_thr=0.01,
- loss_loc=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_shape=dict(type='BoundedIoULoss', beta=0.2, loss_weight=1.0),
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- ga_assigner=dict(
- type='ApproxMaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- ga_sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=-1,
- pos_weight=-1,
- center_ratio=0.2,
- ignore_ratio=0.5,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_label=False),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='proposal_fast')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-# runner configs
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/ga_rpn_r50_caffe_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/guided_anchoring/ga_rpn_r50_caffe_fpn_1x_coco.py b/configs/guided_anchoring/ga_rpn_r50_caffe_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..56a2e343ad89ab1c402e95a4ad3fd3020db7f081
--- /dev/null
+++ b/configs/guided_anchoring/ga_rpn_r50_caffe_fpn_1x_coco.py
@@ -0,0 +1,46 @@
+_base_ = '../rpn/rpn_r50_caffe_fpn_1x_coco.py'
+model = dict(
+ rpn_head=dict(
+ _delete_=True,
+ type='GARPNHead',
+ in_channels=256,
+ feat_channels=256,
+ octave_base_scale=8,
+ scales_per_octave=3,
+ octave_ratios=[0.5, 1.0, 2.0],
+ anchor_strides=[4, 8, 16, 32, 64],
+ anchor_base_sizes=None,
+ anchoring_means=[.0, .0, .0, .0],
+ anchoring_stds=[0.07, 0.07, 0.14, 0.14],
+ target_means=(.0, .0, .0, .0),
+ target_stds=[0.07, 0.07, 0.11, 0.11],
+ loc_filter_thr=0.01,
+ loss_loc=dict(
+ type='FocalLoss',
+ use_sigmoid=True,
+ gamma=2.0,
+ alpha=0.25,
+ loss_weight=1.0),
+ loss_shape=dict(type='BoundedIoULoss', beta=0.2, loss_weight=1.0),
+ loss_cls=dict(
+ type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
+ loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
+# model training and testing settings
+train_cfg = dict(
+ rpn=dict(
+ ga_assigner=dict(
+ type='ApproxMaxIoUAssigner',
+ pos_iou_thr=0.7,
+ neg_iou_thr=0.3,
+ min_pos_iou=0.3,
+ ignore_iof_thr=-1),
+ ga_sampler=dict(
+ type='RandomSampler',
+ num=256,
+ pos_fraction=0.5,
+ neg_pos_ub=-1,
+ add_gt_as_proposals=False),
+ allowed_border=-1,
+ center_ratio=0.2,
+ ignore_ratio=0.5))
+work_dir = './work_dirs/ga_rpn_r50_caffe_fpn_1x'
diff --git a/configs/guided_anchoring/ga_rpn_r50_fpn_1x_coco.py b/configs/guided_anchoring/ga_rpn_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..682a7a7b73a1c19a8ea0cb37816047478868bdc1
--- /dev/null
+++ b/configs/guided_anchoring/ga_rpn_r50_fpn_1x_coco.py
@@ -0,0 +1,46 @@
+_base_ = '../rpn/rpn_r50_fpn_1x_coco.py'
+model = dict(
+ rpn_head=dict(
+ _delete_=True,
+ type='GARPNHead',
+ in_channels=256,
+ feat_channels=256,
+ octave_base_scale=8,
+ scales_per_octave=3,
+ octave_ratios=[0.5, 1.0, 2.0],
+ anchor_strides=[4, 8, 16, 32, 64],
+ anchor_base_sizes=None,
+ anchoring_means=[.0, .0, .0, .0],
+ anchoring_stds=[0.07, 0.07, 0.14, 0.14],
+ target_means=(.0, .0, .0, .0),
+ target_stds=[0.07, 0.07, 0.11, 0.11],
+ loc_filter_thr=0.01,
+ loss_loc=dict(
+ type='FocalLoss',
+ use_sigmoid=True,
+ gamma=2.0,
+ alpha=0.25,
+ loss_weight=1.0),
+ loss_shape=dict(type='BoundedIoULoss', beta=0.2, loss_weight=1.0),
+ loss_cls=dict(
+ type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
+ loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
+# model training and testing settings
+train_cfg = dict(
+ rpn=dict(
+ ga_assigner=dict(
+ type='ApproxMaxIoUAssigner',
+ pos_iou_thr=0.7,
+ neg_iou_thr=0.3,
+ min_pos_iou=0.3,
+ ignore_iof_thr=-1),
+ ga_sampler=dict(
+ type='RandomSampler',
+ num=256,
+ pos_fraction=0.5,
+ neg_pos_ub=-1,
+ add_gt_as_proposals=False),
+ allowed_border=-1,
+ center_ratio=0.2,
+ ignore_ratio=0.5))
+work_dir = './work_dirs/ga_rpn_r50_fpn_1x'
diff --git a/configs/guided_anchoring/ga_rpn_x101_32x4d_fpn_1x.py b/configs/guided_anchoring/ga_rpn_x101_32x4d_fpn_1x.py
deleted file mode 100644
index a8b00113abcef48c7c54b7f9101febacdf7a2dd0..0000000000000000000000000000000000000000
--- a/configs/guided_anchoring/ga_rpn_x101_32x4d_fpn_1x.py
+++ /dev/null
@@ -1,159 +0,0 @@
-# model settings
-model = dict(
- type='RPN',
- pretrained='open-mmlab://resnext101_32x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=32,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='GARPNHead',
- in_channels=256,
- feat_channels=256,
- octave_base_scale=8,
- scales_per_octave=3,
- octave_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- anchor_base_sizes=None,
- anchoring_means=[.0, .0, .0, .0],
- anchoring_stds=[0.07, 0.07, 0.14, 0.14],
- target_means=(.0, .0, .0, .0),
- target_stds=[0.07, 0.07, 0.11, 0.11],
- loc_filter_thr=0.01,
- loss_loc=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_shape=dict(type='BoundedIoULoss', beta=0.2, loss_weight=1.0),
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- ga_assigner=dict(
- type='ApproxMaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- ga_sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=-1,
- pos_weight=-1,
- center_ratio=0.2,
- ignore_ratio=0.5,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_label=False),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='proposal_fast')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-# runner configs
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/ga_rpn_x101_32x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/guided_anchoring/ga_rpn_x101_32x4d_fpn_1x_coco.py b/configs/guided_anchoring/ga_rpn_x101_32x4d_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..a03dc61261e1a84a7d3df2bcec38f62074d6f7e4
--- /dev/null
+++ b/configs/guided_anchoring/ga_rpn_x101_32x4d_fpn_1x_coco.py
@@ -0,0 +1,14 @@
+_base_ = './ga_rpn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_32x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=32,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+work_dir = './work_dirs/ga_rpn_x101_32x4d_fpn_1x'
diff --git a/configs/hrnet/cascade_mask_rcnn_hrnetv2p_w32_20e.py b/configs/hrnet/cascade_mask_rcnn_hrnetv2p_w32_20e.py
deleted file mode 100644
index feb96f579b6cba90a2c7a0e2b1325e6cd559b72b..0000000000000000000000000000000000000000
--- a/configs/hrnet/cascade_mask_rcnn_hrnetv2p_w32_20e.py
+++ /dev/null
@@ -1,269 +0,0 @@
-# model settings
-model = dict(
- type='CascadeRCNN',
- num_stages=3,
- pretrained='open-mmlab://msra/hrnetv2_w32',
- backbone=dict(
- type='HRNet',
- extra=dict(
- stage1=dict(
- num_modules=1,
- num_branches=1,
- block='BOTTLENECK',
- num_blocks=(4, ),
- num_channels=(64, )),
- stage2=dict(
- num_modules=1,
- num_branches=2,
- block='BASIC',
- num_blocks=(4, 4),
- num_channels=(32, 64)),
- stage3=dict(
- num_modules=4,
- num_branches=3,
- block='BASIC',
- num_blocks=(4, 4, 4),
- num_channels=(32, 64, 128)),
- stage4=dict(
- num_modules=3,
- num_branches=4,
- block='BASIC',
- num_blocks=(4, 4, 4, 4),
- num_channels=(32, 64, 128, 256)))),
- neck=dict(type='HRFPN', in_channels=[32, 64, 128, 256], out_channels=256),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=[
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
- ],
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 19])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 20
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/cascade_mask_rcnn_hrnetv2p_w32_20e'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/hrnet/cascade_mask_rcnn_hrnetv2p_w32_20e_coco.py b/configs/hrnet/cascade_mask_rcnn_hrnetv2p_w32_20e_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..6886e844f6ee02f4a8bfa7327bd66b4b55106028
--- /dev/null
+++ b/configs/hrnet/cascade_mask_rcnn_hrnetv2p_w32_20e_coco.py
@@ -0,0 +1,40 @@
+_base_ = '../cascade_rcnn/cascade_mask_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://msra/hrnetv2_w32',
+ backbone=dict(
+ _delete_=True,
+ type='HRNet',
+ extra=dict(
+ stage1=dict(
+ num_modules=1,
+ num_branches=1,
+ block='BOTTLENECK',
+ num_blocks=(4, ),
+ num_channels=(64, )),
+ stage2=dict(
+ num_modules=1,
+ num_branches=2,
+ block='BASIC',
+ num_blocks=(4, 4),
+ num_channels=(32, 64)),
+ stage3=dict(
+ num_modules=4,
+ num_branches=3,
+ block='BASIC',
+ num_blocks=(4, 4, 4),
+ num_channels=(32, 64, 128)),
+ stage4=dict(
+ num_modules=3,
+ num_branches=4,
+ block='BASIC',
+ num_blocks=(4, 4, 4, 4),
+ num_channels=(32, 64, 128, 256)))),
+ neck=dict(
+ _delete_=True,
+ type='HRFPN',
+ in_channels=[32, 64, 128, 256],
+ out_channels=256))
+# learning policy
+lr_config = dict(step=[16, 19])
+total_epochs = 20
+work_dir = './work_dirs/cascade_mask_rcnn_hrnetv2p_w32_20e'
diff --git a/configs/hrnet/cascade_rcnn_hrnetv2p_w32_20e.py b/configs/hrnet/cascade_rcnn_hrnetv2p_w32_20e.py
deleted file mode 100644
index 7fc85b80e9262519cedab83f540fba615149464b..0000000000000000000000000000000000000000
--- a/configs/hrnet/cascade_rcnn_hrnetv2p_w32_20e.py
+++ /dev/null
@@ -1,250 +0,0 @@
-# model settings
-model = dict(
- type='CascadeRCNN',
- num_stages=3,
- pretrained='open-mmlab://msra/hrnetv2_w32',
- backbone=dict(
- type='HRNet',
- extra=dict(
- stage1=dict(
- num_modules=1,
- num_branches=1,
- block='BOTTLENECK',
- num_blocks=(4, ),
- num_channels=(64, )),
- stage2=dict(
- num_modules=1,
- num_branches=2,
- block='BASIC',
- num_blocks=(4, 4),
- num_channels=(32, 64)),
- stage3=dict(
- num_modules=4,
- num_branches=3,
- block='BASIC',
- num_blocks=(4, 4, 4),
- num_channels=(32, 64, 128)),
- stage4=dict(
- num_modules=3,
- num_branches=4,
- block='BASIC',
- num_blocks=(4, 4, 4, 4),
- num_channels=(32, 64, 128, 256)))),
- neck=dict(type='HRFPN', in_channels=[32, 64, 128, 256], out_channels=256),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=[
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- ])
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 19])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 20
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/cascade_rcnn_hrnetv2p_w32'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/hrnet/cascade_rcnn_hrnetv2p_w32_20e_coco.py b/configs/hrnet/cascade_rcnn_hrnetv2p_w32_20e_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..f0b4791393abfdcf117e11d095a49ff23a5093f5
--- /dev/null
+++ b/configs/hrnet/cascade_rcnn_hrnetv2p_w32_20e_coco.py
@@ -0,0 +1,40 @@
+_base_ = '../cascade_rcnn/cascade_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://msra/hrnetv2_w32',
+ backbone=dict(
+ _delete_=True,
+ type='HRNet',
+ extra=dict(
+ stage1=dict(
+ num_modules=1,
+ num_branches=1,
+ block='BOTTLENECK',
+ num_blocks=(4, ),
+ num_channels=(64, )),
+ stage2=dict(
+ num_modules=1,
+ num_branches=2,
+ block='BASIC',
+ num_blocks=(4, 4),
+ num_channels=(32, 64)),
+ stage3=dict(
+ num_modules=4,
+ num_branches=3,
+ block='BASIC',
+ num_blocks=(4, 4, 4),
+ num_channels=(32, 64, 128)),
+ stage4=dict(
+ num_modules=3,
+ num_branches=4,
+ block='BASIC',
+ num_blocks=(4, 4, 4, 4),
+ num_channels=(32, 64, 128, 256)))),
+ neck=dict(
+ _delete_=True,
+ type='HRFPN',
+ in_channels=[32, 64, 128, 256],
+ out_channels=256))
+# learning policy
+lr_config = dict(step=[16, 19])
+total_epochs = 20
+work_dir = './work_dirs/cascade_rcnn_hrnetv2p_w32'
diff --git a/configs/hrnet/faster_rcnn_hrnetv2p_w18_1x.py b/configs/hrnet/faster_rcnn_hrnetv2p_w18_1x.py
deleted file mode 100644
index 44a0c46e62f8acb231b1ec35da47c7d5e5570b8c..0000000000000000000000000000000000000000
--- a/configs/hrnet/faster_rcnn_hrnetv2p_w18_1x.py
+++ /dev/null
@@ -1,191 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='open-mmlab://msra/hrnetv2_w18',
- backbone=dict(
- type='HRNet',
- extra=dict(
- stage1=dict(
- num_modules=1,
- num_branches=1,
- block='BOTTLENECK',
- num_blocks=(4, ),
- num_channels=(64, )),
- stage2=dict(
- num_modules=1,
- num_branches=2,
- block='BASIC',
- num_blocks=(4, 4),
- num_channels=(18, 36)),
- stage3=dict(
- num_modules=4,
- num_branches=3,
- block='BASIC',
- num_blocks=(4, 4, 4),
- num_channels=(18, 36, 72)),
- stage4=dict(
- num_modules=3,
- num_branches=4,
- block='BASIC',
- num_blocks=(4, 4, 4, 4),
- num_channels=(18, 36, 72, 144)))),
- neck=dict(type='HRFPN', in_channels=[18, 36, 72, 144], out_channels=256),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_hrnetv2p_w18_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/hrnet/faster_rcnn_hrnetv2p_w18_1x_coco.py b/configs/hrnet/faster_rcnn_hrnetv2p_w18_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..58de31a63d218e99ec38f4653aedba978c12ea4c
--- /dev/null
+++ b/configs/hrnet/faster_rcnn_hrnetv2p_w18_1x_coco.py
@@ -0,0 +1,11 @@
+_base_ = './faster_rcnn_hrnetv2p_w32_1x_coco.py'
+# model settings
+model = dict(
+ pretrained='open-mmlab://msra/hrnetv2_w18',
+ backbone=dict(
+ extra=dict(
+ stage2=dict(num_channels=(18, 36)),
+ stage3=dict(num_channels=(18, 36, 72)),
+ stage4=dict(num_channels=(18, 36, 72, 144)))),
+ neck=dict(type='HRFPN', in_channels=[18, 36, 72, 144], out_channels=256))
+work_dir = './work_dirs/faster_rcnn_hrnetv2p_w18_1x'
diff --git a/configs/hrnet/faster_rcnn_hrnetv2p_w32_1x.py b/configs/hrnet/faster_rcnn_hrnetv2p_w32_1x.py
deleted file mode 100644
index 820dda100fcaa108418444ff16025e2b55447597..0000000000000000000000000000000000000000
--- a/configs/hrnet/faster_rcnn_hrnetv2p_w32_1x.py
+++ /dev/null
@@ -1,191 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='open-mmlab://msra/hrnetv2_w32',
- backbone=dict(
- type='HRNet',
- extra=dict(
- stage1=dict(
- num_modules=1,
- num_branches=1,
- block='BOTTLENECK',
- num_blocks=(4, ),
- num_channels=(64, )),
- stage2=dict(
- num_modules=1,
- num_branches=2,
- block='BASIC',
- num_blocks=(4, 4),
- num_channels=(32, 64)),
- stage3=dict(
- num_modules=4,
- num_branches=3,
- block='BASIC',
- num_blocks=(4, 4, 4),
- num_channels=(32, 64, 128)),
- stage4=dict(
- num_modules=3,
- num_branches=4,
- block='BASIC',
- num_blocks=(4, 4, 4, 4),
- num_channels=(32, 64, 128, 256)))),
- neck=dict(type='HRFPN', in_channels=[32, 64, 128, 256], out_channels=256),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_hrnetv2p_w32_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/hrnet/faster_rcnn_hrnetv2p_w32_1x_coco.py b/configs/hrnet/faster_rcnn_hrnetv2p_w32_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..cb56ca63bdc8b1621390f45e3ac776733a51f4d8
--- /dev/null
+++ b/configs/hrnet/faster_rcnn_hrnetv2p_w32_1x_coco.py
@@ -0,0 +1,37 @@
+_base_ = '../faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://msra/hrnetv2_w32',
+ backbone=dict(
+ _delete_=True,
+ type='HRNet',
+ extra=dict(
+ stage1=dict(
+ num_modules=1,
+ num_branches=1,
+ block='BOTTLENECK',
+ num_blocks=(4, ),
+ num_channels=(64, )),
+ stage2=dict(
+ num_modules=1,
+ num_branches=2,
+ block='BASIC',
+ num_blocks=(4, 4),
+ num_channels=(32, 64)),
+ stage3=dict(
+ num_modules=4,
+ num_branches=3,
+ block='BASIC',
+ num_blocks=(4, 4, 4),
+ num_channels=(32, 64, 128)),
+ stage4=dict(
+ num_modules=3,
+ num_branches=4,
+ block='BASIC',
+ num_blocks=(4, 4, 4, 4),
+ num_channels=(32, 64, 128, 256)))),
+ neck=dict(
+ _delete_=True,
+ type='HRFPN',
+ in_channels=[32, 64, 128, 256],
+ out_channels=256))
+work_dir = './work_dirs/faster_rcnn_hrnetv2p_w32_1x'
diff --git a/configs/hrnet/faster_rcnn_hrnetv2p_w40_1x.py b/configs/hrnet/faster_rcnn_hrnetv2p_w40_1x.py
deleted file mode 100644
index 6a9bcf03b8b497e7ce39b49e46a29fcb91bfd3f7..0000000000000000000000000000000000000000
--- a/configs/hrnet/faster_rcnn_hrnetv2p_w40_1x.py
+++ /dev/null
@@ -1,191 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='open-mmlab://msra/hrnetv2_w40',
- backbone=dict(
- type='HRNet',
- extra=dict(
- stage1=dict(
- num_modules=1,
- num_branches=1,
- block='BOTTLENECK',
- num_blocks=(4, ),
- num_channels=(64, )),
- stage2=dict(
- num_modules=1,
- num_branches=2,
- block='BASIC',
- num_blocks=(4, 4),
- num_channels=(40, 80)),
- stage3=dict(
- num_modules=4,
- num_branches=3,
- block='BASIC',
- num_blocks=(4, 4, 4),
- num_channels=(40, 80, 160)),
- stage4=dict(
- num_modules=3,
- num_branches=4,
- block='BASIC',
- num_blocks=(4, 4, 4, 4),
- num_channels=(40, 80, 160, 320)))),
- neck=dict(type='HRFPN', in_channels=[40, 80, 160, 320], out_channels=256),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/faster_rcnn_hrnetv2p_w40_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/hrnet/faster_rcnn_hrnetv2p_w40_1x_coco.py b/configs/hrnet/faster_rcnn_hrnetv2p_w40_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..3445cc6bd2f034648a07f2fc6afb7e226c8f4e07
--- /dev/null
+++ b/configs/hrnet/faster_rcnn_hrnetv2p_w40_1x_coco.py
@@ -0,0 +1,11 @@
+_base_ = './faster_rcnn_hrnetv2p_w32_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://msra/hrnetv2_w40',
+ backbone=dict(
+ type='HRNet',
+ extra=dict(
+ stage2=dict(num_channels=(40, 80)),
+ stage3=dict(num_channels=(40, 80, 160)),
+ stage4=dict(num_channels=(40, 80, 160, 320)))),
+ neck=dict(type='HRFPN', in_channels=[40, 80, 160, 320], out_channels=256))
+work_dir = './work_dirs/faster_rcnn_hrnetv2p_w40_1x'
diff --git a/configs/hrnet/fcos_hrnetv2p_w32_gn-head_4x4_1x_coco.py b/configs/hrnet/fcos_hrnetv2p_w32_gn-head_4x4_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..33e1ce33538fe4f24ffbfff251844f7af8e57388
--- /dev/null
+++ b/configs/hrnet/fcos_hrnetv2p_w32_gn-head_4x4_1x_coco.py
@@ -0,0 +1,39 @@
+_base_ = '../fcos/fcos_r50_caffe_fpn_gn-head_4x4_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://msra/hrnetv2_w32',
+ backbone=dict(
+ _delete_=True,
+ type='HRNet',
+ extra=dict(
+ stage1=dict(
+ num_modules=1,
+ num_branches=1,
+ block='BOTTLENECK',
+ num_blocks=(4, ),
+ num_channels=(64, )),
+ stage2=dict(
+ num_modules=1,
+ num_branches=2,
+ block='BASIC',
+ num_blocks=(4, 4),
+ num_channels=(32, 64)),
+ stage3=dict(
+ num_modules=4,
+ num_branches=3,
+ block='BASIC',
+ num_blocks=(4, 4, 4),
+ num_channels=(32, 64, 128)),
+ stage4=dict(
+ num_modules=3,
+ num_branches=4,
+ block='BASIC',
+ num_blocks=(4, 4, 4, 4),
+ num_channels=(32, 64, 128, 256)))),
+ neck=dict(
+ _delete_=True,
+ type='HRFPN',
+ in_channels=[32, 64, 128, 256],
+ out_channels=256,
+ stride=2,
+ num_outs=5))
+work_dir = './work_dirs/fcos_hrnetv2p_w32_gn_1x_4gpu'
diff --git a/configs/hrnet/fcos_hrnetv2p_w32_gn_1x_4gpu.py b/configs/hrnet/fcos_hrnetv2p_w32_gn_1x_4gpu.py
deleted file mode 100644
index e2c52250c96095d328380112f56234b3df5ec7dd..0000000000000000000000000000000000000000
--- a/configs/hrnet/fcos_hrnetv2p_w32_gn_1x_4gpu.py
+++ /dev/null
@@ -1,151 +0,0 @@
-# model settings
-model = dict(
- type='FCOS',
- pretrained='open-mmlab://msra/hrnetv2_w32',
- backbone=dict(
- type='HRNet',
- extra=dict(
- stage1=dict(
- num_modules=1,
- num_branches=1,
- block='BOTTLENECK',
- num_blocks=(4, ),
- num_channels=(64, )),
- stage2=dict(
- num_modules=1,
- num_branches=2,
- block='BASIC',
- num_blocks=(4, 4),
- num_channels=(32, 64)),
- stage3=dict(
- num_modules=4,
- num_branches=3,
- block='BASIC',
- num_blocks=(4, 4, 4),
- num_channels=(32, 64, 128)),
- stage4=dict(
- num_modules=3,
- num_branches=4,
- block='BASIC',
- num_blocks=(4, 4, 4, 4),
- num_channels=(32, 64, 128, 256)))),
- neck=dict(
- type='HRFPN',
- in_channels=[32, 64, 128, 256],
- out_channels=256,
- stride=2,
- num_outs=5),
- bbox_head=dict(
- type='FCOSHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- strides=[8, 16, 32, 64, 128],
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox=dict(type='IoULoss', loss_weight=1.0),
- loss_centerness=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=4,
- workers_per_gpu=4,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(
- type='SGD',
- lr=0.01,
- momentum=0.9,
- weight_decay=0.0001,
- paramwise_options=dict(bias_lr_mult=2., bias_decay_mult=0.))
-optimizer_config = dict(grad_clip=None)
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='constant',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/fcos_hrnetv2p_w32_gn_1x_4gpu'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/hrnet/htc_hrnetv2p_w32_20e.py b/configs/hrnet/htc_hrnetv2p_w32_20e.py
deleted file mode 100644
index 40836c592bfcf2aa80f018703fad0842d0bfd088..0000000000000000000000000000000000000000
--- a/configs/hrnet/htc_hrnetv2p_w32_20e.py
+++ /dev/null
@@ -1,310 +0,0 @@
-# model settings
-model = dict(
- type='HybridTaskCascade',
- num_stages=3,
- pretrained='open-mmlab://msra/hrnetv2_w32',
- interleaved=True,
- mask_info_flow=True,
- backbone=dict(
- type='HRNet',
- extra=dict(
- stage1=dict(
- num_modules=1,
- num_branches=1,
- block='BOTTLENECK',
- num_blocks=(4, ),
- num_channels=(64, )),
- stage2=dict(
- num_modules=1,
- num_branches=2,
- block='BASIC',
- num_blocks=(4, 4),
- num_channels=(32, 64)),
- stage3=dict(
- num_modules=4,
- num_branches=3,
- block='BASIC',
- num_blocks=(4, 4, 4),
- num_channels=(32, 64, 128)),
- stage4=dict(
- num_modules=3,
- num_branches=4,
- block='BASIC',
- num_blocks=(4, 4, 4, 4),
- num_channels=(32, 64, 128, 256)))),
- neck=dict(type='HRFPN', in_channels=[32, 64, 128, 256], out_channels=256),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=[
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
- ],
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=[
- dict(
- type='HTCMaskHead',
- with_conv_res=False,
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)),
- dict(
- type='HTCMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)),
- dict(
- type='HTCMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0))
- ],
- semantic_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[8]),
- semantic_head=dict(
- type='FusedSemanticHead',
- num_ins=5,
- fusion_level=1,
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=183,
- ignore_label=255,
- loss_weight=0.2))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.001,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='LoadAnnotations', with_bbox=True, with_mask=True, with_seg=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='SegRescale', scale_factor=1 / 8),
- dict(type='DefaultFormatBundle'),
- dict(
- type='Collect',
- keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks', 'gt_semantic_seg']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- seg_prefix=data_root + 'stuffthingmaps/train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 19])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 20
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/htc_hrnetv2p_w32_20e'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/hrnet/htc_hrnetv2p_w32_20e_coco.py b/configs/hrnet/htc_hrnetv2p_w32_20e_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..eba65a8e63ffc552de144fb774b4021c9a6c5751
--- /dev/null
+++ b/configs/hrnet/htc_hrnetv2p_w32_20e_coco.py
@@ -0,0 +1,37 @@
+_base_ = '../htc/htc_r50_fpn_20e_coco.py'
+model = dict(
+ pretrained='open-mmlab://msra/hrnetv2_w32',
+ backbone=dict(
+ _delete_=True,
+ type='HRNet',
+ extra=dict(
+ stage1=dict(
+ num_modules=1,
+ num_branches=1,
+ block='BOTTLENECK',
+ num_blocks=(4, ),
+ num_channels=(64, )),
+ stage2=dict(
+ num_modules=1,
+ num_branches=2,
+ block='BASIC',
+ num_blocks=(4, 4),
+ num_channels=(32, 64)),
+ stage3=dict(
+ num_modules=4,
+ num_branches=3,
+ block='BASIC',
+ num_blocks=(4, 4, 4),
+ num_channels=(32, 64, 128)),
+ stage4=dict(
+ num_modules=3,
+ num_branches=4,
+ block='BASIC',
+ num_blocks=(4, 4, 4, 4),
+ num_channels=(32, 64, 128, 256)))),
+ neck=dict(
+ _delete_=True,
+ type='HRFPN',
+ in_channels=[32, 64, 128, 256],
+ out_channels=256))
+work_dir = './work_dirs/htc_hrnetv2p_w32_20e'
diff --git a/configs/hrnet/mask_rcnn_hrnetv2p_w18_1x.py b/configs/hrnet/mask_rcnn_hrnetv2p_w18_1x.py
deleted file mode 100644
index 0e562ffb4af7f640be06a9036e15cd457884ac3d..0000000000000000000000000000000000000000
--- a/configs/hrnet/mask_rcnn_hrnetv2p_w18_1x.py
+++ /dev/null
@@ -1,206 +0,0 @@
-# model settings
-model = dict(
- type='MaskRCNN',
- pretrained='open-mmlab://msra/hrnetv2_w18',
- backbone=dict(
- type='HRNet',
- extra=dict(
- stage1=dict(
- num_modules=1,
- num_branches=1,
- block='BOTTLENECK',
- num_blocks=(4, ),
- num_channels=(64, )),
- stage2=dict(
- num_modules=1,
- num_branches=2,
- block='BASIC',
- num_blocks=(4, 4),
- num_channels=(18, 36)),
- stage3=dict(
- num_modules=4,
- num_branches=3,
- block='BASIC',
- num_blocks=(4, 4, 4),
- num_channels=(18, 36, 72)),
- stage4=dict(
- num_modules=3,
- num_branches=4,
- block='BASIC',
- num_blocks=(4, 4, 4, 4),
- num_channels=(18, 36, 72, 144)))),
- neck=dict(type='HRFPN', in_channels=[18, 36, 72, 144], out_channels=256),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-# if you use 8 GPUs for training, please change lr to 0.02
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_hrnetv2p_w18_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/hrnet/mask_rcnn_hrnetv2p_w18_1x_coco.py b/configs/hrnet/mask_rcnn_hrnetv2p_w18_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..f52f1efbe0f73639c86f81ec36a31ba6053d1e5f
--- /dev/null
+++ b/configs/hrnet/mask_rcnn_hrnetv2p_w18_1x_coco.py
@@ -0,0 +1,10 @@
+_base_ = './mask_rcnn_hrnetv2p_w32_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://msra/hrnetv2_w18',
+ backbone=dict(
+ extra=dict(
+ stage2=dict(num_channels=(18, 36)),
+ stage3=dict(num_channels=(18, 36, 72)),
+ stage4=dict(num_channels=(18, 36, 72, 144)))),
+ neck=dict(type='HRFPN', in_channels=[18, 36, 72, 144], out_channels=256))
+work_dir = './work_dirs/mask_rcnn_hrnetv2p_w18_1x'
diff --git a/configs/hrnet/mask_rcnn_hrnetv2p_w32_1x.py b/configs/hrnet/mask_rcnn_hrnetv2p_w32_1x.py
deleted file mode 100644
index 4511426f437866a0d92c4513c8da5dc334196e1d..0000000000000000000000000000000000000000
--- a/configs/hrnet/mask_rcnn_hrnetv2p_w32_1x.py
+++ /dev/null
@@ -1,205 +0,0 @@
-# model settings
-model = dict(
- type='MaskRCNN',
- pretrained='open-mmlab://msra/hrnetv2_w32',
- backbone=dict(
- type='HRNet',
- extra=dict(
- stage1=dict(
- num_modules=1,
- num_branches=1,
- block='BOTTLENECK',
- num_blocks=(4, ),
- num_channels=(64, )),
- stage2=dict(
- num_modules=1,
- num_branches=2,
- block='BASIC',
- num_blocks=(4, 4),
- num_channels=(32, 64)),
- stage3=dict(
- num_modules=4,
- num_branches=3,
- block='BASIC',
- num_blocks=(4, 4, 4),
- num_channels=(32, 64, 128)),
- stage4=dict(
- num_modules=3,
- num_branches=4,
- block='BASIC',
- num_blocks=(4, 4, 4, 4),
- num_channels=(32, 64, 128, 256)))),
- neck=dict(type='HRFPN', in_channels=[32, 64, 128, 256], out_channels=256),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_hrnetv2p_w32_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/hrnet/mask_rcnn_hrnetv2p_w32_1x_coco.py b/configs/hrnet/mask_rcnn_hrnetv2p_w32_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..7abeeda7a179851103324507fef9dfb7b7765880
--- /dev/null
+++ b/configs/hrnet/mask_rcnn_hrnetv2p_w32_1x_coco.py
@@ -0,0 +1,37 @@
+_base_ = '../mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://msra/hrnetv2_w32',
+ backbone=dict(
+ _delete_=True,
+ type='HRNet',
+ extra=dict(
+ stage1=dict(
+ num_modules=1,
+ num_branches=1,
+ block='BOTTLENECK',
+ num_blocks=(4, ),
+ num_channels=(64, )),
+ stage2=dict(
+ num_modules=1,
+ num_branches=2,
+ block='BASIC',
+ num_blocks=(4, 4),
+ num_channels=(32, 64)),
+ stage3=dict(
+ num_modules=4,
+ num_branches=3,
+ block='BASIC',
+ num_blocks=(4, 4, 4),
+ num_channels=(32, 64, 128)),
+ stage4=dict(
+ num_modules=3,
+ num_branches=4,
+ block='BASIC',
+ num_blocks=(4, 4, 4, 4),
+ num_channels=(32, 64, 128, 256)))),
+ neck=dict(
+ _delete_=True,
+ type='HRFPN',
+ in_channels=[32, 64, 128, 256],
+ out_channels=256))
+work_dir = './work_dirs/mask_rcnn_hrnetv2p_w32_1x'
diff --git a/configs/htc/htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e.py b/configs/htc/htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e.py
deleted file mode 100644
index ab1afd859c2bcddbb31d80f33e96222086f867c3..0000000000000000000000000000000000000000
--- a/configs/htc/htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e.py
+++ /dev/null
@@ -1,303 +0,0 @@
-# model settings
-model = dict(
- type='HybridTaskCascade',
- num_stages=3,
- pretrained='open-mmlab://resnext101_64x4d',
- interleaved=True,
- mask_info_flow=True,
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=64,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
- stage_with_dcn=(False, True, True, True)),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=[
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
- ],
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=[
- dict(
- type='HTCMaskHead',
- with_conv_res=False,
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)),
- dict(
- type='HTCMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)),
- dict(
- type='HTCMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0))
- ],
- semantic_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[8]),
- semantic_head=dict(
- type='FusedSemanticHead',
- num_ins=5,
- fusion_level=1,
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=183,
- ignore_label=255,
- loss_weight=0.2))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.001,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='LoadAnnotations', with_bbox=True, with_mask=True, with_seg=True),
- dict(
- type='Resize',
- img_scale=[(1600, 400), (1600, 1400)],
- multiscale_mode='range',
- keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='SegRescale', scale_factor=1 / 8),
- dict(type='DefaultFormatBundle'),
- dict(
- type='Collect',
- keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks', 'gt_semantic_seg']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=1,
- workers_per_gpu=1,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- seg_prefix=data_root + 'stuffthingmaps/train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 19])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 20
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/htc/htc_r101_fpn_20e.py b/configs/htc/htc_r101_fpn_20e.py
deleted file mode 100644
index 5b46ab53b1ffc8a1bcf039f0b3527ee514f59c9a..0000000000000000000000000000000000000000
--- a/configs/htc/htc_r101_fpn_20e.py
+++ /dev/null
@@ -1,295 +0,0 @@
-# model settings
-model = dict(
- type='HybridTaskCascade',
- num_stages=3,
- pretrained='torchvision://resnet101',
- interleaved=True,
- mask_info_flow=True,
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=[
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
- ],
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=[
- dict(
- type='HTCMaskHead',
- with_conv_res=False,
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)),
- dict(
- type='HTCMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)),
- dict(
- type='HTCMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0))
- ],
- semantic_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[8]),
- semantic_head=dict(
- type='FusedSemanticHead',
- num_ins=5,
- fusion_level=1,
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=183,
- ignore_label=255,
- loss_weight=0.2))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.001,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='LoadAnnotations', with_bbox=True, with_mask=True, with_seg=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='SegRescale', scale_factor=1 / 8),
- dict(type='DefaultFormatBundle'),
- dict(
- type='Collect',
- keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks', 'gt_semantic_seg']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- seg_prefix=data_root + 'stuffthingmaps/train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 19])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 20
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/htc_r101_fpn_20e'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/htc/htc_r101_fpn_20e_coco.py b/configs/htc/htc_r101_fpn_20e_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..fa34f13317456ac2934e1757241bf7871b1364b4
--- /dev/null
+++ b/configs/htc/htc_r101_fpn_20e_coco.py
@@ -0,0 +1,6 @@
+_base_ = './htc_r50_fpn_1x_coco.py'
+model = dict(pretrained='torchvision://resnet101', backbone=dict(depth=101))
+# learning policy
+lr_config = dict(step=[16, 19])
+total_epochs = 20
+work_dir = './work_dirs/htc_r101_fpn_20e'
diff --git a/configs/htc/htc_r50_fpn_1x.py b/configs/htc/htc_r50_fpn_1x.py
deleted file mode 100644
index f28d7df24f8d009a1c5b56e2988a545cff917ce1..0000000000000000000000000000000000000000
--- a/configs/htc/htc_r50_fpn_1x.py
+++ /dev/null
@@ -1,295 +0,0 @@
-# model settings
-model = dict(
- type='HybridTaskCascade',
- num_stages=3,
- pretrained='torchvision://resnet50',
- interleaved=True,
- mask_info_flow=True,
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=[
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
- ],
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=[
- dict(
- type='HTCMaskHead',
- with_conv_res=False,
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)),
- dict(
- type='HTCMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)),
- dict(
- type='HTCMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0))
- ],
- semantic_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[8]),
- semantic_head=dict(
- type='FusedSemanticHead',
- num_ins=5,
- fusion_level=1,
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=183,
- ignore_label=255,
- loss_weight=0.2))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.001,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='LoadAnnotations', with_bbox=True, with_mask=True, with_seg=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='SegRescale', scale_factor=1 / 8),
- dict(type='DefaultFormatBundle'),
- dict(
- type='Collect',
- keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks', 'gt_semantic_seg']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- seg_prefix=data_root + 'stuffthingmaps/train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/htc_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/htc/htc_r50_fpn_1x_coco.py b/configs/htc/htc_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..6da8e5fc206e693e1b88c4baaef7096470751ccf
--- /dev/null
+++ b/configs/htc/htc_r50_fpn_1x_coco.py
@@ -0,0 +1,56 @@
+_base_ = './htc_without_semantic_r50_fpn_1x_coco.py'
+model = dict(
+ semantic_roi_extractor=dict(
+ type='SingleRoIExtractor',
+ roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
+ out_channels=256,
+ featmap_strides=[8]),
+ semantic_head=dict(
+ type='FusedSemanticHead',
+ num_ins=5,
+ fusion_level=1,
+ num_convs=4,
+ in_channels=256,
+ conv_out_channels=256,
+ num_classes=183,
+ ignore_label=255,
+ loss_weight=0.2))
+data_root = 'data/coco/'
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='LoadAnnotations', with_bbox=True, with_mask=True, with_seg=True),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='SegRescale', scale_factor=1 / 8),
+ dict(type='DefaultFormatBundle'),
+ dict(
+ type='Collect',
+ keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks', 'gt_semantic_seg']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(1333, 800),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ train=dict(
+ seg_prefix=data_root + 'stuffthingmaps/train2017/',
+ pipeline=train_pipeline),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+work_dir = './work_dirs/htc_r50_fpn_1x'
diff --git a/configs/htc/htc_r50_fpn_20e.py b/configs/htc/htc_r50_fpn_20e.py
deleted file mode 100644
index ec596b9ac089002ed5c85614102bf0dd1e6de24c..0000000000000000000000000000000000000000
--- a/configs/htc/htc_r50_fpn_20e.py
+++ /dev/null
@@ -1,295 +0,0 @@
-# model settings
-model = dict(
- type='HybridTaskCascade',
- num_stages=3,
- pretrained='torchvision://resnet50',
- interleaved=True,
- mask_info_flow=True,
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=[
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
- ],
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=[
- dict(
- type='HTCMaskHead',
- with_conv_res=False,
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)),
- dict(
- type='HTCMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)),
- dict(
- type='HTCMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0))
- ],
- semantic_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[8]),
- semantic_head=dict(
- type='FusedSemanticHead',
- num_ins=5,
- fusion_level=1,
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=183,
- ignore_label=255,
- loss_weight=0.2))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.001,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='LoadAnnotations', with_bbox=True, with_mask=True, with_seg=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='SegRescale', scale_factor=1 / 8),
- dict(type='DefaultFormatBundle'),
- dict(
- type='Collect',
- keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks', 'gt_semantic_seg']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- seg_prefix=data_root + 'stuffthingmaps/train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 19])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 20
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/htc_r50_fpn_20e'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/htc/htc_r50_fpn_20e_coco.py b/configs/htc/htc_r50_fpn_20e_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..00947bf35d9d4f1e31382404c7cbf6cb3b252b96
--- /dev/null
+++ b/configs/htc/htc_r50_fpn_20e_coco.py
@@ -0,0 +1,5 @@
+_base_ = './htc_r50_fpn_1x_coco.py'
+# learning policy
+lr_config = dict(step=[16, 19])
+total_epochs = 20
+work_dir = './work_dirs/htc_r50_fpn_20e'
diff --git a/configs/htc/htc_without_semantic_r50_fpn_1x.py b/configs/htc/htc_without_semantic_r50_fpn_1x_coco.py
similarity index 77%
rename from configs/htc/htc_without_semantic_r50_fpn_1x.py
rename to configs/htc/htc_without_semantic_r50_fpn_1x_coco.py
index 6616e8e202f07c9135eab0c780f8642fe700ccea..ec0bbce44c1ed3683a48c176f580493daefa0a8a 100644
--- a/configs/htc/htc_without_semantic_r50_fpn_1x.py
+++ b/configs/htc/htc_without_semantic_r50_fpn_1x_coco.py
@@ -1,3 +1,7 @@
+_base_ = [
+ '../_base_/datasets/coco_instance.py',
+ '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
+]
# model settings
model = dict(
type='HybridTaskCascade',
@@ -37,8 +41,7 @@ model = dict(
featmap_strides=[4, 8, 16, 32]),
bbox_head=[
dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
+ type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
@@ -50,8 +53,7 @@ model = dict(
type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
+ type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
@@ -63,8 +65,7 @@ model = dict(
type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
+ type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
@@ -197,21 +198,8 @@ test_cfg = dict(
nms=dict(type='nms', iou_thr=0.5),
max_per_img=100,
mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
img_norm_cfg = dict(
mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
test_pipeline = [
dict(type='LoadImageFromFile'),
dict(
@@ -228,48 +216,5 @@ test_pipeline = [
])
]
data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
+ val=dict(pipeline=test_pipeline), test=dict(pipeline=test_pipeline))
work_dir = './work_dirs/htc_without_semantic_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/htc/htc_x101_32x4d_fpn_16x1_20e_coco.py b/configs/htc/htc_x101_32x4d_fpn_16x1_20e_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..cbb32558dea4230b23cdb14c47fa687a21dd66fc
--- /dev/null
+++ b/configs/htc/htc_x101_32x4d_fpn_16x1_20e_coco.py
@@ -0,0 +1,18 @@
+_base_ = './htc_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_32x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=32,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+data = dict(imgs_per_gpu=1, workers_per_gpu=1)
+# learning policy
+lr_config = dict(step=[16, 19])
+total_epochs = 20
+work_dir = './work_dirs/htc_x101_32x4d_fpn_20e'
diff --git a/configs/htc/htc_x101_32x4d_fpn_20e_16gpu.py b/configs/htc/htc_x101_32x4d_fpn_20e_16gpu.py
deleted file mode 100644
index 65c94a661bb2d676a3494150b26a4d7d2dfdeadc..0000000000000000000000000000000000000000
--- a/configs/htc/htc_x101_32x4d_fpn_20e_16gpu.py
+++ /dev/null
@@ -1,297 +0,0 @@
-# model settings
-model = dict(
- type='HybridTaskCascade',
- num_stages=3,
- pretrained='open-mmlab://resnext101_32x4d',
- interleaved=True,
- mask_info_flow=True,
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=32,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=[
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
- ],
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=[
- dict(
- type='HTCMaskHead',
- with_conv_res=False,
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)),
- dict(
- type='HTCMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)),
- dict(
- type='HTCMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0))
- ],
- semantic_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[8]),
- semantic_head=dict(
- type='FusedSemanticHead',
- num_ins=5,
- fusion_level=1,
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=183,
- ignore_label=255,
- loss_weight=0.2))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.001,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='LoadAnnotations', with_bbox=True, with_mask=True, with_seg=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='SegRescale', scale_factor=1 / 8),
- dict(type='DefaultFormatBundle'),
- dict(
- type='Collect',
- keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks', 'gt_semantic_seg']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=1,
- workers_per_gpu=1,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- seg_prefix=data_root + 'stuffthingmaps/train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 19])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 20
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/htc_x101_32x4d_fpn_20e'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/htc/htc_x101_64x4d_fpn_16x1_20e_coco.py b/configs/htc/htc_x101_64x4d_fpn_16x1_20e_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..797d219d3b79453413c2b71ec73759bf097db7bd
--- /dev/null
+++ b/configs/htc/htc_x101_64x4d_fpn_16x1_20e_coco.py
@@ -0,0 +1,18 @@
+_base_ = './htc_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_64x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=64,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+data = dict(imgs_per_gpu=1, workers_per_gpu=1)
+# learning policy
+lr_config = dict(step=[16, 19])
+total_epochs = 20
+work_dir = './work_dirs/htc_x101_64x4d_fpn_20e'
diff --git a/configs/htc/htc_x101_64x4d_fpn_20e_16gpu.py b/configs/htc/htc_x101_64x4d_fpn_20e_16gpu.py
deleted file mode 100644
index 574736b5741f9ba18e53e8709ec471b764f8acea..0000000000000000000000000000000000000000
--- a/configs/htc/htc_x101_64x4d_fpn_20e_16gpu.py
+++ /dev/null
@@ -1,297 +0,0 @@
-# model settings
-model = dict(
- type='HybridTaskCascade',
- num_stages=3,
- pretrained='open-mmlab://resnext101_64x4d',
- interleaved=True,
- mask_info_flow=True,
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=64,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=[
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
- ],
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=[
- dict(
- type='HTCMaskHead',
- with_conv_res=False,
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)),
- dict(
- type='HTCMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)),
- dict(
- type='HTCMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0))
- ],
- semantic_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[8]),
- semantic_head=dict(
- type='FusedSemanticHead',
- num_ins=5,
- fusion_level=1,
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=183,
- ignore_label=255,
- loss_weight=0.2))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.001,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='LoadAnnotations', with_bbox=True, with_mask=True, with_seg=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='SegRescale', scale_factor=1 / 8),
- dict(type='DefaultFormatBundle'),
- dict(
- type='Collect',
- keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks', 'gt_semantic_seg']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=1,
- workers_per_gpu=1,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- seg_prefix=data_root + 'stuffthingmaps/train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 19])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 20
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/htc_x101_64x4d_fpn_20e'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/htc/htc_x101_64x4d_fpn_dconv_c3-c5_mstrain_400_1400_16x1_20e_coco.py b/configs/htc/htc_x101_64x4d_fpn_dconv_c3-c5_mstrain_400_1400_16x1_20e_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..45027582dd68a6ca1e6dd79a7e213559853868dd
--- /dev/null
+++ b/configs/htc/htc_x101_64x4d_fpn_dconv_c3-c5_mstrain_400_1400_16x1_20e_coco.py
@@ -0,0 +1,42 @@
+_base_ = './htc_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_64x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=64,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch',
+ dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
+ stage_with_dcn=(False, True, True, True)))
+# dataset settings
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='LoadAnnotations', with_bbox=True, with_mask=True, with_seg=True),
+ dict(
+ type='Resize',
+ img_scale=[(1600, 400), (1600, 1400)],
+ multiscale_mode='range',
+ keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='SegRescale', scale_factor=1 / 8),
+ dict(type='DefaultFormatBundle'),
+ dict(
+ type='Collect',
+ keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks', 'gt_semantic_seg']),
+]
+data = dict(
+ imgs_per_gpu=1, workers_per_gpu=1, train=dict(pipeline=train_pipeline))
+# learning policy
+lr_config = dict(step=[16, 19])
+total_epochs = 20
+work_dir = './work_dirs/htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e'
diff --git a/configs/instaboost/cascade_mask_rcnn_r50_fpn_instaboost_4x.py b/configs/instaboost/cascade_mask_rcnn_r50_fpn_instaboost_4x.py
deleted file mode 100644
index 4ce744df3f375a288085ab4f48e13039df5a24b5..0000000000000000000000000000000000000000
--- a/configs/instaboost/cascade_mask_rcnn_r50_fpn_instaboost_4x.py
+++ /dev/null
@@ -1,265 +0,0 @@
-# model settings
-model = dict(
- type='CascadeRCNN',
- num_stages=3,
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=[
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.05, 0.05, 0.1, 0.1],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.033, 0.033, 0.067, 0.067],
- reg_class_agnostic=True,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
- ],
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=[
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.6,
- neg_iou_thr=0.6,
- min_pos_iou=0.6,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False),
- dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.7,
- min_pos_iou=0.7,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False)
- ],
- stage_loss_weights=[1, 0.5, 0.25])
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='InstaBoost',
- action_candidate=('normal', 'horizontal', 'skip'),
- action_prob=(1, 0, 0),
- scale=(0.8, 1.2),
- dx=15,
- dy=15,
- theta=(-1, 1),
- color_prob=0.5,
- hflag=False,
- aug_ratio=0.5),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[32, 44])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 48
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/cascade_mask_rcnn_r50_fpn_instaboost_4x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/instaboost/cascade_mask_rcnn_r50_fpn_instaboost_4x_coco.py b/configs/instaboost/cascade_mask_rcnn_r50_fpn_instaboost_4x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..80ae5bee3031cbc652fc79bd12c1ac138323ebad
--- /dev/null
+++ b/configs/instaboost/cascade_mask_rcnn_r50_fpn_instaboost_4x_coco.py
@@ -0,0 +1,29 @@
+_base_ = '../cascade_rcnn/cascade_mask_rcnn_r50_fpn_1x_coco.py'
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='InstaBoost',
+ action_candidate=('normal', 'horizontal', 'skip'),
+ action_prob=(1, 0, 0),
+ scale=(0.8, 1.2),
+ dx=15,
+ dy=15,
+ theta=(-1, 1),
+ color_prob=0.5,
+ hflag=False,
+ aug_ratio=0.5),
+ dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
+]
+data = dict(train=dict(pipeline=train_pipeline))
+# learning policy
+lr_config = dict(step=[32, 44])
+total_epochs = 48
+work_dir = './work_dirs/cascade_mask_rcnn_r50_fpn_instaboost_4x'
diff --git a/configs/instaboost/mask_rcnn_r50_fpn_instaboost_4x.py b/configs/instaboost/mask_rcnn_r50_fpn_instaboost_4x.py
deleted file mode 100644
index 169652dc7cc5e0c37010a234a669293306fcd078..0000000000000000000000000000000000000000
--- a/configs/instaboost/mask_rcnn_r50_fpn_instaboost_4x.py
+++ /dev/null
@@ -1,201 +0,0 @@
-# model settings
-model = dict(
- type='MaskRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='InstaBoost',
- action_candidate=('normal', 'horizontal', 'skip'),
- action_prob=(1, 0, 0),
- scale=(0.8, 1.2),
- dx=15,
- dy=15,
- theta=(-1, 1),
- color_prob=0.5,
- hflag=False,
- aug_ratio=0.5),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[32, 44])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 48
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_r50_fpn_instaboost_4x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/instaboost/mask_rcnn_r50_fpn_instaboost_4x_coco.py b/configs/instaboost/mask_rcnn_r50_fpn_instaboost_4x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..8df3596225e00e26d7e17cd1029caf805c1790d1
--- /dev/null
+++ b/configs/instaboost/mask_rcnn_r50_fpn_instaboost_4x_coco.py
@@ -0,0 +1,29 @@
+_base_ = '../mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py'
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='InstaBoost',
+ action_candidate=('normal', 'horizontal', 'skip'),
+ action_prob=(1, 0, 0),
+ scale=(0.8, 1.2),
+ dx=15,
+ dy=15,
+ theta=(-1, 1),
+ color_prob=0.5,
+ hflag=False,
+ aug_ratio=0.5),
+ dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
+]
+data = dict(train=dict(pipeline=train_pipeline))
+# learning policy
+lr_config = dict(step=[32, 44])
+total_epochs = 48
+work_dir = './work_dirs/mask_rcnn_r50_fpn_instaboost_4x'
diff --git a/configs/instaboost/ssd300_coco_instaboost_4x.py b/configs/instaboost/ssd300_coco_instaboost_4x.py
deleted file mode 100644
index 3c7de8700b37b298ee3608f0471340c457afada5..0000000000000000000000000000000000000000
--- a/configs/instaboost/ssd300_coco_instaboost_4x.py
+++ /dev/null
@@ -1,146 +0,0 @@
-# model settings
-input_size = 300
-model = dict(
- type='SingleStageDetector',
- pretrained='open-mmlab://vgg16_caffe',
- backbone=dict(
- type='SSDVGG',
- input_size=input_size,
- depth=16,
- with_last_pool=False,
- ceil_mode=True,
- out_indices=(3, 4),
- out_feature_indices=(22, 34),
- l2_norm_scale=20),
- neck=None,
- bbox_head=dict(
- type='SSDHead',
- input_size=input_size,
- in_channels=(512, 1024, 512, 256, 256, 256),
- num_classes=81,
- anchor_strides=(8, 16, 32, 64, 100, 300),
- basesize_ratio_range=(0.15, 0.9),
- anchor_ratios=([2], [2, 3], [2, 3], [2, 3], [2], [2]),
- target_means=(.0, .0, .0, .0),
- target_stds=(0.1, 0.1, 0.2, 0.2)))
-cudnn_benchmark = True
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.,
- ignore_iof_thr=-1,
- gt_max_assign_all=False),
- smoothl1_beta=1.,
- allowed_border=-1,
- pos_weight=-1,
- neg_pos_ratio=3,
- debug=False)
-test_cfg = dict(
- nms=dict(type='nms', iou_thr=0.45),
- min_bbox_size=0,
- score_thr=0.02,
- max_per_img=200)
-# model training and testing settings
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(mean=[123.675, 116.28, 103.53], std=[1, 1, 1], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile', to_float32=True),
- dict(
- type='InstaBoost',
- action_candidate=('normal', 'horizontal', 'skip'),
- action_prob=(1, 0, 0),
- scale=(0.8, 1.2),
- dx=15,
- dy=15,
- theta=(-1, 1),
- color_prob=0.5,
- hflag=False,
- aug_ratio=0.5),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(
- type='PhotoMetricDistortion',
- brightness_delta=32,
- contrast_range=(0.5, 1.5),
- saturation_range=(0.5, 1.5),
- hue_delta=18),
- dict(
- type='Expand',
- mean=img_norm_cfg['mean'],
- to_rgb=img_norm_cfg['to_rgb'],
- ratio_range=(1, 4)),
- dict(
- type='MinIoURandomCrop',
- min_ious=(0.1, 0.3, 0.5, 0.7, 0.9),
- min_crop_size=0.3),
- dict(type='Resize', img_scale=(300, 300), keep_ratio=False),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(300, 300),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=False),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=8,
- workers_per_gpu=3,
- train=dict(
- type='RepeatDataset',
- times=5,
- dataset=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline)),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=2e-3, momentum=0.9, weight_decay=5e-4)
-optimizer_config = dict()
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[64, 88])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 96
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/ssd300_coco_instaboost_4x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/instaboost/ssd300_coco_instaboost_4x_coco.py b/configs/instaboost/ssd300_coco_instaboost_4x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..472d18341e034221bcc247ad148d98f21f22607e
--- /dev/null
+++ b/configs/instaboost/ssd300_coco_instaboost_4x_coco.py
@@ -0,0 +1,42 @@
+_base_ = '../ssd/ssd300_coco.py'
+img_norm_cfg = dict(mean=[123.675, 116.28, 103.53], std=[1, 1, 1], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile', to_float32=True),
+ dict(
+ type='InstaBoost',
+ action_candidate=('normal', 'horizontal', 'skip'),
+ action_prob=(1, 0, 0),
+ scale=(0.8, 1.2),
+ dx=15,
+ dy=15,
+ theta=(-1, 1),
+ color_prob=0.5,
+ hflag=False,
+ aug_ratio=0.5),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(
+ type='PhotoMetricDistortion',
+ brightness_delta=32,
+ contrast_range=(0.5, 1.5),
+ saturation_range=(0.5, 1.5),
+ hue_delta=18),
+ dict(
+ type='Expand',
+ mean=img_norm_cfg['mean'],
+ to_rgb=img_norm_cfg['to_rgb'],
+ ratio_range=(1, 4)),
+ dict(
+ type='MinIoURandomCrop',
+ min_ious=(0.1, 0.3, 0.5, 0.7, 0.9),
+ min_crop_size=0.3),
+ dict(type='Resize', img_scale=(300, 300), keep_ratio=False),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+data = dict(train=dict(dataset=dict(pipeline=train_pipeline)))
+# learning policy
+lr_config = dict(step=[64, 88])
+total_epochs = 96
+work_dir = './work_dirs/ssd300_coco_instaboost_4x'
diff --git a/configs/libra_rcnn/libra_fast_rcnn_r50_fpn_1x.py b/configs/libra_rcnn/libra_fast_rcnn_r50_fpn_1x.py
deleted file mode 100644
index 839225268c8a591eb337c65686b84a6f319b062d..0000000000000000000000000000000000000000
--- a/configs/libra_rcnn/libra_fast_rcnn_r50_fpn_1x.py
+++ /dev/null
@@ -1,155 +0,0 @@
-# model settings
-model = dict(
- type='FastRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=[
- dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- dict(
- type='BFP',
- in_channels=256,
- num_levels=5,
- refine_level=2,
- refine_type='non_local')
- ],
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(
- type='BalancedL1Loss',
- alpha=0.5,
- gamma=1.5,
- beta=1.0,
- loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='CombinedSampler',
- num=512,
- pos_fraction=0.25,
- add_gt_as_proposals=True,
- pos_sampler=dict(type='InstanceBalancedPosSampler'),
- neg_sampler=dict(
- type='IoUBalancedNegSampler',
- floor_thr=-1,
- floor_fraction=0,
- num_bins=3)),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadProposals', num_max_proposals=2000),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'proposals', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadProposals', num_max_proposals=None),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img', 'proposals']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- proposal_file=data_root +
- 'libra_proposals/rpn_r50_fpn_1x_train2017.pkl',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- proposal_file=data_root + 'libra_proposals/rpn_r50_fpn_1x_val2017.pkl',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- proposal_file=data_root + 'libra_proposals/rpn_r50_fpn_1x_val2017.pkl',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/libra_fast_rcnn_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/libra_rcnn/libra_fast_rcnn_r50_fpn_1x_coco.py b/configs/libra_rcnn/libra_fast_rcnn_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..440fe9785400da644c9219120b13639079faf00b
--- /dev/null
+++ b/configs/libra_rcnn/libra_fast_rcnn_r50_fpn_1x_coco.py
@@ -0,0 +1,50 @@
+_base_ = '../fast_rcnn/fast_rcnn_r50_fpn_1x_coco.py'
+# model settings
+model = dict(
+ neck=[
+ dict(
+ type='FPN',
+ in_channels=[256, 512, 1024, 2048],
+ out_channels=256,
+ num_outs=5),
+ dict(
+ type='BFP',
+ in_channels=256,
+ num_levels=5,
+ refine_level=2,
+ refine_type='non_local')
+ ],
+ bbox_head=dict(
+ loss_bbox=dict(
+ _delete_=True,
+ type='BalancedL1Loss',
+ alpha=0.5,
+ gamma=1.5,
+ beta=1.0,
+ loss_weight=1.0)))
+# model training and testing settings
+train_cfg = dict(
+ rcnn=dict(
+ sampler=dict(
+ _delete_=True,
+ type='CombinedSampler',
+ num=512,
+ pos_fraction=0.25,
+ add_gt_as_proposals=True,
+ pos_sampler=dict(type='InstanceBalancedPosSampler'),
+ neg_sampler=dict(
+ type='IoUBalancedNegSampler',
+ floor_thr=-1,
+ floor_fraction=0,
+ num_bins=3))))
+# dataset settings
+dataset_type = 'CocoDataset'
+data_root = 'data/coco/'
+data = dict(
+ train=dict(proposal_file=data_root +
+ 'libra_proposals/rpn_r50_fpn_1x_train2017.pkl'),
+ val=dict(proposal_file=data_root +
+ 'libra_proposals/rpn_r50_fpn_1x_val2017.pkl'),
+ test=dict(proposal_file=data_root +
+ 'libra_proposals/rpn_r50_fpn_1x_val2017.pkl'))
+work_dir = './work_dirs/libra_fast_rcnn_r50_fpn_1x'
diff --git a/configs/libra_rcnn/libra_faster_rcnn_r101_fpn_1x.py b/configs/libra_rcnn/libra_faster_rcnn_r101_fpn_1x.py
deleted file mode 100644
index af3a0b6749bfbf2d77ed0459b3a49c8f4aefb2fe..0000000000000000000000000000000000000000
--- a/configs/libra_rcnn/libra_faster_rcnn_r101_fpn_1x.py
+++ /dev/null
@@ -1,194 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='torchvision://resnet101',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=[
- dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- dict(
- type='BFP',
- in_channels=256,
- num_levels=5,
- refine_level=2,
- refine_type='non_local')
- ],
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(
- type='BalancedL1Loss',
- alpha=0.5,
- gamma=1.5,
- beta=1.0,
- loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=5,
- add_gt_as_proposals=False),
- allowed_border=-1,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='CombinedSampler',
- num=512,
- pos_fraction=0.25,
- add_gt_as_proposals=True,
- pos_sampler=dict(type='InstanceBalancedPosSampler'),
- neg_sampler=dict(
- type='IoUBalancedNegSampler',
- floor_thr=-1,
- floor_fraction=0,
- num_bins=3)),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/libra_faster_rcnn_r101_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/libra_rcnn/libra_faster_rcnn_r101_fpn_1x_coco.py b/configs/libra_rcnn/libra_faster_rcnn_r101_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..6b4c15a28899758e5371649fb728f6db46f91028
--- /dev/null
+++ b/configs/libra_rcnn/libra_faster_rcnn_r101_fpn_1x_coco.py
@@ -0,0 +1,3 @@
+_base_ = './libra_faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(pretrained='torchvision://resnet101', backbone=dict(depth=101))
+work_dir = './work_dirs/libra_faster_rcnn_r101_fpn_1x'
diff --git a/configs/libra_rcnn/libra_faster_rcnn_r50_fpn_1x.py b/configs/libra_rcnn/libra_faster_rcnn_r50_fpn_1x.py
deleted file mode 100644
index 41de0c62259ffc79584d3ef1e443c5e4e945c70d..0000000000000000000000000000000000000000
--- a/configs/libra_rcnn/libra_faster_rcnn_r50_fpn_1x.py
+++ /dev/null
@@ -1,194 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=[
- dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- dict(
- type='BFP',
- in_channels=256,
- num_levels=5,
- refine_level=2,
- refine_type='non_local')
- ],
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(
- type='BalancedL1Loss',
- alpha=0.5,
- gamma=1.5,
- beta=1.0,
- loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=5,
- add_gt_as_proposals=False),
- allowed_border=-1,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='CombinedSampler',
- num=512,
- pos_fraction=0.25,
- add_gt_as_proposals=True,
- pos_sampler=dict(type='InstanceBalancedPosSampler'),
- neg_sampler=dict(
- type='IoUBalancedNegSampler',
- floor_thr=-1,
- floor_fraction=0,
- num_bins=3)),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/libra_faster_rcnn_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/libra_rcnn/libra_faster_rcnn_r50_fpn_1x_coco.py b/configs/libra_rcnn/libra_faster_rcnn_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..5e99672974d80d237e8bfea19f7243cc7724adb7
--- /dev/null
+++ b/configs/libra_rcnn/libra_faster_rcnn_r50_fpn_1x_coco.py
@@ -0,0 +1,41 @@
+_base_ = '../faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
+# model settings
+model = dict(
+ neck=[
+ dict(
+ type='FPN',
+ in_channels=[256, 512, 1024, 2048],
+ out_channels=256,
+ num_outs=5),
+ dict(
+ type='BFP',
+ in_channels=256,
+ num_levels=5,
+ refine_level=2,
+ refine_type='non_local')
+ ],
+ bbox_head=dict(
+ loss_bbox=dict(
+ _delete_=True,
+ type='BalancedL1Loss',
+ alpha=0.5,
+ gamma=1.5,
+ beta=1.0,
+ loss_weight=1.0)))
+# model training and testing settings
+train_cfg = dict(
+ rpn=dict(sampler=dict(neg_pos_ub=5), allowed_border=-1),
+ rcnn=dict(
+ sampler=dict(
+ _delete_=True,
+ type='CombinedSampler',
+ num=512,
+ pos_fraction=0.25,
+ add_gt_as_proposals=True,
+ pos_sampler=dict(type='InstanceBalancedPosSampler'),
+ neg_sampler=dict(
+ type='IoUBalancedNegSampler',
+ floor_thr=-1,
+ floor_fraction=0,
+ num_bins=3))))
+work_dir = './work_dirs/libra_faster_rcnn_r50_fpn_1x'
diff --git a/configs/libra_rcnn/libra_faster_rcnn_x101_64x4d_fpn_1x.py b/configs/libra_rcnn/libra_faster_rcnn_x101_64x4d_fpn_1x.py
deleted file mode 100644
index 50ec1df825ac0d853c02a08a7636ceb589e5c556..0000000000000000000000000000000000000000
--- a/configs/libra_rcnn/libra_faster_rcnn_x101_64x4d_fpn_1x.py
+++ /dev/null
@@ -1,196 +0,0 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='open-mmlab://resnext101_64x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=64,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=[
- dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- dict(
- type='BFP',
- in_channels=256,
- num_levels=5,
- refine_level=2,
- refine_type='non_local')
- ],
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(
- type='BalancedL1Loss',
- alpha=0.5,
- gamma=1.5,
- beta=1.0,
- loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=5,
- add_gt_as_proposals=False),
- allowed_border=-1,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='CombinedSampler',
- num=512,
- pos_fraction=0.25,
- add_gt_as_proposals=True,
- pos_sampler=dict(type='InstanceBalancedPosSampler'),
- neg_sampler=dict(
- type='IoUBalancedNegSampler',
- floor_thr=-1,
- floor_fraction=0,
- num_bins=3)),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/libra_faster_rcnn_x101_64x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/libra_rcnn/libra_faster_rcnn_x101_64x4d_fpn_1x_coco.py b/configs/libra_rcnn/libra_faster_rcnn_x101_64x4d_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..87674040644f45e7a49676e48a063d4ef338346a
--- /dev/null
+++ b/configs/libra_rcnn/libra_faster_rcnn_x101_64x4d_fpn_1x_coco.py
@@ -0,0 +1,14 @@
+_base_ = './libra_faster_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_64x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=64,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+work_dir = './work_dirs/libra_faster_rcnn_x101_64x4d_fpn_1x'
diff --git a/configs/libra_rcnn/libra_retinanet_r50_fpn_1x.py b/configs/libra_rcnn/libra_retinanet_r50_fpn_1x.py
deleted file mode 100644
index 543ec4a99d6868a9f1041b80641bc758252388fd..0000000000000000000000000000000000000000
--- a/configs/libra_rcnn/libra_retinanet_r50_fpn_1x.py
+++ /dev/null
@@ -1,144 +0,0 @@
-# model settings
-model = dict(
- type='RetinaNet',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=[
- dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5),
- dict(
- type='BFP',
- in_channels=256,
- num_levels=5,
- refine_level=1,
- refine_type='non_local')
- ],
- bbox_head=dict(
- type='RetinaHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- octave_base_scale=4,
- scales_per_octave=3,
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[8, 16, 32, 64, 128],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox=dict(
- type='BalancedL1Loss',
- alpha=0.5,
- gamma=1.5,
- beta=0.11,
- loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/libra_retinanet_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/libra_rcnn/libra_retinanet_r50_fpn_1x_coco.py b/configs/libra_rcnn/libra_retinanet_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..34f6222701eacec140f04e85ce14f6922494f4b6
--- /dev/null
+++ b/configs/libra_rcnn/libra_retinanet_r50_fpn_1x_coco.py
@@ -0,0 +1,27 @@
+_base_ = '../retinanet/retinanet_r50_fpn_1x_coco.py'
+# model settings
+model = dict(
+ neck=[
+ dict(
+ type='FPN',
+ in_channels=[256, 512, 1024, 2048],
+ out_channels=256,
+ start_level=1,
+ add_extra_convs=True,
+ num_outs=5),
+ dict(
+ type='BFP',
+ in_channels=256,
+ num_levels=5,
+ refine_level=1,
+ refine_type='non_local')
+ ],
+ bbox_head=dict(
+ loss_bbox=dict(
+ _delete_=True,
+ type='BalancedL1Loss',
+ alpha=0.5,
+ gamma=1.5,
+ beta=0.11,
+ loss_weight=1.0)))
+work_dir = './work_dirs/libra_retinanet_r50_fpn_1x'
diff --git a/configs/mask_rcnn/mask_rcnn_r101_fpn_1x_coco.py b/configs/mask_rcnn/mask_rcnn_r101_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..92efafc29e5b258dc1163a0801bd025f1464925d
--- /dev/null
+++ b/configs/mask_rcnn/mask_rcnn_r101_fpn_1x_coco.py
@@ -0,0 +1,3 @@
+_base_ = './mask_rcnn_r50_fpn_1x_coco.py'
+model = dict(pretrained='torchvision://resnet101', backbone=dict(depth=101))
+work_dir = './work_dirs/mask_rcnn_r101_fpn_1x'
diff --git a/configs/mask_rcnn/mask_rcnn_r50_caffe_c4_1x_coco.py b/configs/mask_rcnn/mask_rcnn_r50_caffe_c4_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..16a4fd648751149fb25992d1fc400a52e54ca15c
--- /dev/null
+++ b/configs/mask_rcnn/mask_rcnn_r50_caffe_c4_1x_coco.py
@@ -0,0 +1,40 @@
+_base_ = [
+ '../_base_/models/mask_rcnn_r50_caffe_c4.py',
+ '../_base_/datasets/coco_instance.py',
+ '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
+]
+# use caffe img_norm
+img_norm_cfg = dict(
+ mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(1333, 800),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ train=dict(pipeline=train_pipeline),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+# optimizer
+optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
+work_dir = './work_dirs/mask_rcnn_r50_caffe_c4_1x'
diff --git a/configs/mask_rcnn/mask_rcnn_r50_caffe_fpn_1x_coco.py b/configs/mask_rcnn/mask_rcnn_r50_caffe_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..5e8a6d653f4b354d2e6e28acba5441aef653dbfd
--- /dev/null
+++ b/configs/mask_rcnn/mask_rcnn_r50_caffe_fpn_1x_coco.py
@@ -0,0 +1,37 @@
+_base_ = './mask_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnet50_caffe',
+ backbone=dict(norm_cfg=dict(requires_grad=False), style='caffe'))
+# use caffe img_norm
+img_norm_cfg = dict(
+ mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(1333, 800),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ train=dict(pipeline=train_pipeline),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+work_dir = './work_dirs/mask_rcnn_r50_caffe_fpn_1x'
diff --git a/configs/mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py b/configs/mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..6d84957d2aff0fb8da863c5d034520ebf036268f
--- /dev/null
+++ b/configs/mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py
@@ -0,0 +1,6 @@
+_base_ = [
+ '../_base_/models/mask_rcnn_r50_fpn.py',
+ '../_base_/datasets/coco_instance.py',
+ '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
+]
+work_dir = './work_dirs/mask_rcnn_r50_fpn_1x'
diff --git a/configs/mask_rcnn/mask_rcnn_x101_32x4d_fpn_1x_coco.py b/configs/mask_rcnn/mask_rcnn_x101_32x4d_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..3092a71c042a81e52118935c2dfaaae8bd7d0279
--- /dev/null
+++ b/configs/mask_rcnn/mask_rcnn_x101_32x4d_fpn_1x_coco.py
@@ -0,0 +1,14 @@
+_base_ = './mask_rcnn_r101_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_32x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=32,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+work_dir = './work_dirs/mask_rcnn_x101_32x4d_fpn_1x'
diff --git a/configs/mask_rcnn/mask_rcnn_x101_64x4d_fpn_1x_coco.py b/configs/mask_rcnn/mask_rcnn_x101_64x4d_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..35a56b07d67d4bf2f86b3b000130a3fd7d26b7a9
--- /dev/null
+++ b/configs/mask_rcnn/mask_rcnn_x101_64x4d_fpn_1x_coco.py
@@ -0,0 +1,14 @@
+_base_ = './mask_rcnn_x101_32x4d_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_64x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=64,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+work_dir = './work_dirs/mask_rcnn_x101_64x4d_fpn_1x'
diff --git a/configs/mask_rcnn_r101_fpn_1x.py b/configs/mask_rcnn_r101_fpn_1x.py
deleted file mode 100644
index 47451f09cf664ef786bd70225606f2a2d18c25ae..0000000000000000000000000000000000000000
--- a/configs/mask_rcnn_r101_fpn_1x.py
+++ /dev/null
@@ -1,190 +0,0 @@
-# model settings
-model = dict(
- type='MaskRCNN',
- pretrained='torchvision://resnet101',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_r101_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/mask_rcnn_x101_32x4d_fpn_1x.py b/configs/mask_rcnn_x101_32x4d_fpn_1x.py
deleted file mode 100644
index 0e33ef4ea1b14ab7b3589a7bfe03d03205122901..0000000000000000000000000000000000000000
--- a/configs/mask_rcnn_x101_32x4d_fpn_1x.py
+++ /dev/null
@@ -1,192 +0,0 @@
-# model settings
-model = dict(
- type='MaskRCNN',
- pretrained='open-mmlab://resnext101_32x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=32,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_x101_32x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/mask_rcnn_x101_64x4d_fpn_1x.py b/configs/mask_rcnn_x101_64x4d_fpn_1x.py
deleted file mode 100644
index 02ef2cee3cb14dbadcae00063f383142623c47d9..0000000000000000000000000000000000000000
--- a/configs/mask_rcnn_x101_64x4d_fpn_1x.py
+++ /dev/null
@@ -1,192 +0,0 @@
-# model settings
-model = dict(
- type='MaskRCNN',
- pretrained='open-mmlab://resnext101_64x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=64,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/mask_rcnn_x101_64x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/ms_rcnn/ms_rcnn_r101_caffe_fpn_1x.py b/configs/ms_rcnn/ms_rcnn_r101_caffe_fpn_1x.py
deleted file mode 100644
index c59a51316dfc853a43b8234993393a043b163e90..0000000000000000000000000000000000000000
--- a/configs/ms_rcnn/ms_rcnn_r101_caffe_fpn_1x.py
+++ /dev/null
@@ -1,200 +0,0 @@
-# model settings
-model = dict(
- type='MaskScoringRCNN',
- pretrained='open-mmlab://resnet101_caffe',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=False),
- style='caffe'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)),
- mask_iou_head=dict(
- type='MaskIoUHead',
- num_convs=4,
- num_fcs=2,
- roi_feat_size=14,
- in_channels=256,
- conv_out_channels=256,
- fc_out_channels=1024,
- num_classes=81))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- mask_thr_binary=0.5,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/ms_rcnn_r101_caffe_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/ms_rcnn/ms_rcnn_r101_caffe_fpn_1x_coco.py b/configs/ms_rcnn/ms_rcnn_r101_caffe_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..9f75a0b876da60197eecbb18d829dfd76eca7768
--- /dev/null
+++ b/configs/ms_rcnn/ms_rcnn_r101_caffe_fpn_1x_coco.py
@@ -0,0 +1,4 @@
+_base_ = './ms_rcnn_r50_caffe_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnet101_caffe', backbone=dict(depth=101))
+work_dir = './work_dirs/ms_rcnn_r101_caffe_fpn_1x'
diff --git a/configs/ms_rcnn/ms_rcnn_r50_caffe_fpn_1x.py b/configs/ms_rcnn/ms_rcnn_r50_caffe_fpn_1x.py
deleted file mode 100644
index 3476a01a32c4c112c78e5853ea2e79fe60793ed1..0000000000000000000000000000000000000000
--- a/configs/ms_rcnn/ms_rcnn_r50_caffe_fpn_1x.py
+++ /dev/null
@@ -1,200 +0,0 @@
-# model settings
-model = dict(
- type='MaskScoringRCNN',
- pretrained='open-mmlab://resnet50_caffe',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=False),
- style='caffe'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)),
- mask_iou_head=dict(
- type='MaskIoUHead',
- num_convs=4,
- num_fcs=2,
- roi_feat_size=14,
- in_channels=256,
- conv_out_channels=256,
- fc_out_channels=1024,
- num_classes=81))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- mask_thr_binary=0.5,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/ms_rcnn_r50_caffe_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/ms_rcnn/ms_rcnn_r50_caffe_fpn_1x_coco.py b/configs/ms_rcnn/ms_rcnn_r50_caffe_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..e245892323b8e771bceec825b0aeb69f3e793f45
--- /dev/null
+++ b/configs/ms_rcnn/ms_rcnn_r50_caffe_fpn_1x_coco.py
@@ -0,0 +1,15 @@
+_base_ = '../mask_rcnn/mask_rcnn_r50_caffe_fpn_1x_coco.py'
+model = dict(
+ type='MaskScoringRCNN',
+ mask_iou_head=dict(
+ type='MaskIoUHead',
+ num_convs=4,
+ num_fcs=2,
+ roi_feat_size=14,
+ in_channels=256,
+ conv_out_channels=256,
+ fc_out_channels=1024,
+ num_classes=81))
+# model training and testing settings
+train_cfg = dict(rcnn=dict(mask_thr_binary=0.5))
+work_dir = './work_dirs/ms_rcnn_r50_caffe_fpn_1x'
diff --git a/configs/ms_rcnn/ms_rcnn_r50_fpn_1x_coco.py b/configs/ms_rcnn/ms_rcnn_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..6911ef01601fdc1f489873cbde157959ce0b77dc
--- /dev/null
+++ b/configs/ms_rcnn/ms_rcnn_r50_fpn_1x_coco.py
@@ -0,0 +1,15 @@
+_base_ = '../mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ type='MaskScoringRCNN',
+ mask_iou_head=dict(
+ type='MaskIoUHead',
+ num_convs=4,
+ num_fcs=2,
+ roi_feat_size=14,
+ in_channels=256,
+ conv_out_channels=256,
+ fc_out_channels=1024,
+ num_classes=81))
+# model training and testing settings
+train_cfg = dict(rcnn=dict(mask_thr_binary=0.5))
+work_dir = './work_dirs/ms_rcnn_r50_fpn_1x'
diff --git a/configs/ms_rcnn/ms_rcnn_x101_64x4d_fpn_1x.py b/configs/ms_rcnn/ms_rcnn_x101_64x4d_fpn_1x.py
deleted file mode 100644
index 856a9307015baada7bb6c4a0576ae71d9eea6462..0000000000000000000000000000000000000000
--- a/configs/ms_rcnn/ms_rcnn_x101_64x4d_fpn_1x.py
+++ /dev/null
@@ -1,202 +0,0 @@
-# model settings
-model = dict(
- type='MaskScoringRCNN',
- pretrained='open-mmlab://resnext101_64x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=64,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)),
- mask_iou_head=dict(
- type='MaskIoUHead',
- num_convs=4,
- num_fcs=2,
- roi_feat_size=14,
- in_channels=256,
- conv_out_channels=256,
- fc_out_channels=1024,
- num_classes=81))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- mask_thr_binary=0.5,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/ms_rcnn_x101_64x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/ms_rcnn/ms_rcnn_x101_64x4d_fpn_1x_coco.py b/configs/ms_rcnn/ms_rcnn_x101_64x4d_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..786c08f730fe42fb58e27adfd14388385df7dcc8
--- /dev/null
+++ b/configs/ms_rcnn/ms_rcnn_x101_64x4d_fpn_1x_coco.py
@@ -0,0 +1,14 @@
+_base_ = './ms_rcnn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_64x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=64,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+work_dir = './work_dirs/ms_rcnn_x101_64x4d_fpn_1x'
diff --git a/configs/nas_fpn/retinanet_crop640_r50_fpn_50e.py b/configs/nas_fpn/retinanet_crop640_r50_fpn_50e.py
deleted file mode 100644
index a921c5b8771e90b747f25f07c776a3421a850e12..0000000000000000000000000000000000000000
--- a/configs/nas_fpn/retinanet_crop640_r50_fpn_50e.py
+++ /dev/null
@@ -1,149 +0,0 @@
-cudnn_benchmark = True
-# model settings
-norm_cfg = dict(type='BN', requires_grad=True)
-model = dict(
- type='RetinaNet',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=norm_cfg,
- norm_eval=False,
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- relu_before_extra_convs=True,
- no_norm_on_lateral=True,
- norm_cfg=norm_cfg,
- num_outs=5),
- bbox_head=dict(
- type='RetinaSepBNHead',
- num_classes=81,
- num_ins=5,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- octave_base_scale=4,
- scales_per_octave=3,
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[8, 16, 32, 64, 128],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(
- type='Resize',
- img_scale=(640, 640),
- ratio_range=(0.8, 1.2),
- keep_ratio=True),
- dict(type='RandomCrop', crop_size=(640, 640)),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size=(640, 640)),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(640, 640),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=64),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=8,
- workers_per_gpu=4,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(
- type='SGD',
- lr=0.08,
- momentum=0.9,
- weight_decay=0.0001,
- paramwise_options=dict(norm_decay_mult=0))
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=1000,
- warmup_ratio=0.1,
- step=[30, 40])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 50
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/retinanet_crop640_r50_fpn_50e'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/nas_fpn/retinanet_crop640_r50_nasfpn_50e.py b/configs/nas_fpn/retinanet_crop640_r50_nasfpn_50e.py
deleted file mode 100644
index 8797430428ec394a1ee5e0d6c2100a128da15fd8..0000000000000000000000000000000000000000
--- a/configs/nas_fpn/retinanet_crop640_r50_nasfpn_50e.py
+++ /dev/null
@@ -1,148 +0,0 @@
-cudnn_benchmark = True
-# model settings
-norm_cfg = dict(type='BN', requires_grad=True)
-model = dict(
- type='RetinaNet',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=norm_cfg,
- norm_eval=False,
- style='pytorch'),
- neck=dict(
- type='NASFPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5,
- stack_times=7,
- start_level=1,
- add_extra_convs=True,
- norm_cfg=norm_cfg),
- bbox_head=dict(
- type='RetinaSepBNHead',
- num_classes=81,
- num_ins=5,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- octave_base_scale=4,
- scales_per_octave=3,
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[8, 16, 32, 64, 128],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(
- type='Resize',
- img_scale=(640, 640),
- ratio_range=(0.8, 1.2),
- keep_ratio=True),
- dict(type='RandomCrop', crop_size=(640, 640)),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size=(640, 640)),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(640, 640),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=128),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=8,
- workers_per_gpu=4,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(
- type='SGD',
- lr=0.08,
- momentum=0.9,
- weight_decay=0.0001,
- paramwise_options=dict(norm_decay_mult=0))
-optimizer_config = dict(grad_clip=None)
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=1000,
- warmup_ratio=0.1,
- step=[30, 40])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 50
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/retinanet_crop640_r50_nasfpn_50e'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/nas_fpn/retinanet_r50_fpn_crop640_50e_coco.py b/configs/nas_fpn/retinanet_r50_fpn_crop640_50e_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..069d5a8bbe5babc994ccf04f09db41c1ed9a22bb
--- /dev/null
+++ b/configs/nas_fpn/retinanet_r50_fpn_crop640_50e_coco.py
@@ -0,0 +1,81 @@
+_base_ = [
+ '../_base_/models/retinanet_r50_fpn.py',
+ '../_base_/datasets/coco_detection.py', '../_base_/default_runtime.py'
+]
+cudnn_benchmark = True
+norm_cfg = dict(type='BN', requires_grad=True)
+model = dict(
+ pretrained='torchvision://resnet50',
+ backbone=dict(
+ type='ResNet',
+ depth=50,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=norm_cfg,
+ norm_eval=False,
+ style='pytorch'),
+ neck=dict(
+ relu_before_extra_convs=True,
+ no_norm_on_lateral=True,
+ norm_cfg=norm_cfg),
+ bbox_head=dict(type='RetinaSepBNHead', num_ins=5, norm_cfg=norm_cfg))
+# training and testing settings
+train_cfg = dict(assigner=dict(neg_iou_thr=0.5))
+# dataset settings
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(
+ type='Resize',
+ img_scale=(640, 640),
+ ratio_range=(0.8, 1.2),
+ keep_ratio=True),
+ dict(type='RandomCrop', crop_size=(640, 640)),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size=(640, 640)),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(640, 640),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=64),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ imgs_per_gpu=8,
+ workers_per_gpu=4,
+ train=dict(pipeline=train_pipeline),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+# optimizer
+optimizer = dict(
+ type='SGD',
+ lr=0.08,
+ momentum=0.9,
+ weight_decay=0.0001,
+ paramwise_options=dict(norm_decay_mult=0))
+optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
+# learning policy
+lr_config = dict(
+ policy='step',
+ warmup='linear',
+ warmup_iters=1000,
+ warmup_ratio=0.1,
+ step=[30, 40])
+# runtime settings
+total_epochs = 50
+work_dir = './work_dirs/retinanet_crop640_r50_fpn_50e'
diff --git a/configs/nas_fpn/retinanet_r50_nasfpn_crop640_50e_coco.py b/configs/nas_fpn/retinanet_r50_nasfpn_crop640_50e_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..8f76590ba177c7b250f73068755ed58b68291df4
--- /dev/null
+++ b/configs/nas_fpn/retinanet_r50_nasfpn_crop640_50e_coco.py
@@ -0,0 +1,80 @@
+_base_ = [
+ '../_base_/models/retinanet_r50_fpn.py',
+ '../_base_/datasets/coco_detection.py', '../_base_/default_runtime.py'
+]
+cudnn_benchmark = True
+# model settings
+norm_cfg = dict(type='BN', requires_grad=True)
+model = dict(
+ type='RetinaNet',
+ pretrained='torchvision://resnet50',
+ backbone=dict(
+ type='ResNet',
+ depth=50,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=norm_cfg,
+ norm_eval=False,
+ style='pytorch'),
+ neck=dict(type='NASFPN', stack_times=7, norm_cfg=norm_cfg),
+ bbox_head=dict(type='RetinaSepBNHead', num_ins=5, norm_cfg=norm_cfg))
+# training and testing settings
+train_cfg = dict(assigner=dict(neg_iou_thr=0.5))
+# dataset settings
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(
+ type='Resize',
+ img_scale=(640, 640),
+ ratio_range=(0.8, 1.2),
+ keep_ratio=True),
+ dict(type='RandomCrop', crop_size=(640, 640)),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size=(640, 640)),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(640, 640),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=128),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ imgs_per_gpu=8,
+ workers_per_gpu=4,
+ train=dict(pipeline=train_pipeline),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+# optimizer
+optimizer = dict(
+ type='SGD',
+ lr=0.08,
+ momentum=0.9,
+ weight_decay=0.0001,
+ paramwise_options=dict(norm_decay_mult=0))
+optimizer_config = dict(grad_clip=None)
+# learning policy
+lr_config = dict(
+ policy='step',
+ warmup='linear',
+ warmup_iters=1000,
+ warmup_ratio=0.1,
+ step=[30, 40])
+# runtime settings
+total_epochs = 50
+work_dir = './work_dirs/retinanet_crop640_r50_nasfpn_50e'
diff --git a/configs/pascal_voc/faster_rcnn_r50_fpn_1x_voc0712.py b/configs/pascal_voc/faster_rcnn_r50_fpn_1x_voc0712.py
index 041cd583462f23bbc0fdaa6428799751d950e70e..1e487606867ba563afabddfafade5335ea4e7163 100644
--- a/configs/pascal_voc/faster_rcnn_r50_fpn_1x_voc0712.py
+++ b/configs/pascal_voc/faster_rcnn_r50_fpn_1x_voc0712.py
@@ -1,177 +1,14 @@
-# model settings
-model = dict(
- type='FasterRCNN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='SharedFCBBoxHead',
- num_fcs=2,
- in_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=21,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
- # soft-nms is also supported for rcnn testing
- # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
-)
-# dataset settings
-dataset_type = 'VOCDataset'
-data_root = 'data/VOCdevkit/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1000, 600), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+_base_ = [
+ '../_base_/models/faster_rcnn_r50_fpn.py', '../_base_/datasets/voc0712.py',
+ '../_base_/default_runtime.py'
]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1000, 600),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type='RepeatDataset',
- times=3,
- dataset=dict(
- type=dataset_type,
- ann_file=[
- data_root + 'VOC2007/ImageSets/Main/trainval.txt',
- data_root + 'VOC2012/ImageSets/Main/trainval.txt'
- ],
- img_prefix=[data_root + 'VOC2007/', data_root + 'VOC2012/'],
- pipeline=train_pipeline)),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'VOC2007/ImageSets/Main/test.txt',
- img_prefix=data_root + 'VOC2007/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'VOC2007/ImageSets/Main/test.txt',
- img_prefix=data_root + 'VOC2007/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='mAP')
+model = dict(bbox_head=dict(num_classes=21))
# optimizer
optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
# learning policy
-lr_config = dict(policy='step', step=[3]) # actual epoch = 3 * 3 = 9
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
+# actual epoch = 3 * 3 = 9
+lr_config = dict(policy='step', step=[3])
# runtime settings
total_epochs = 4 # actual epoch = 4 * 3 = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
work_dir = './work_dirs/faster_rcnn_r50_fpn_1x_voc0712'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/pascal_voc/ssd300_voc.py b/configs/pascal_voc/ssd300_voc.py
deleted file mode 100644
index 4589424d3d3924e37dbb8ec564ede41b6225e26b..0000000000000000000000000000000000000000
--- a/configs/pascal_voc/ssd300_voc.py
+++ /dev/null
@@ -1,138 +0,0 @@
-# model settings
-input_size = 300
-model = dict(
- type='SingleStageDetector',
- pretrained='open-mmlab://vgg16_caffe',
- backbone=dict(
- type='SSDVGG',
- input_size=input_size,
- depth=16,
- with_last_pool=False,
- ceil_mode=True,
- out_indices=(3, 4),
- out_feature_indices=(22, 34),
- l2_norm_scale=20),
- neck=None,
- bbox_head=dict(
- type='SSDHead',
- input_size=input_size,
- in_channels=(512, 1024, 512, 256, 256, 256),
- num_classes=21,
- anchor_strides=(8, 16, 32, 64, 100, 300),
- basesize_ratio_range=(0.2, 0.9),
- anchor_ratios=([2], [2, 3], [2, 3], [2, 3], [2], [2]),
- target_means=(.0, .0, .0, .0),
- target_stds=(0.1, 0.1, 0.2, 0.2)))
-# model training and testing settings
-cudnn_benchmark = True
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.,
- ignore_iof_thr=-1,
- gt_max_assign_all=False),
- smoothl1_beta=1.,
- allowed_border=-1,
- pos_weight=-1,
- neg_pos_ratio=3,
- debug=False)
-test_cfg = dict(
- nms=dict(type='nms', iou_thr=0.45),
- min_bbox_size=0,
- score_thr=0.02,
- max_per_img=200)
-# dataset settings
-dataset_type = 'VOCDataset'
-data_root = 'data/VOCdevkit/'
-img_norm_cfg = dict(mean=[123.675, 116.28, 103.53], std=[1, 1, 1], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile', to_float32=True),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(
- type='PhotoMetricDistortion',
- brightness_delta=32,
- contrast_range=(0.5, 1.5),
- saturation_range=(0.5, 1.5),
- hue_delta=18),
- dict(
- type='Expand',
- mean=img_norm_cfg['mean'],
- to_rgb=img_norm_cfg['to_rgb'],
- ratio_range=(1, 4)),
- dict(
- type='MinIoURandomCrop',
- min_ious=(0.1, 0.3, 0.5, 0.7, 0.9),
- min_crop_size=0.3),
- dict(type='Resize', img_scale=(300, 300), keep_ratio=False),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(300, 300),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=False),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=8,
- workers_per_gpu=3,
- train=dict(
- type='RepeatDataset',
- times=10,
- dataset=dict(
- type=dataset_type,
- ann_file=[
- data_root + 'VOC2007/ImageSets/Main/trainval.txt',
- data_root + 'VOC2012/ImageSets/Main/trainval.txt'
- ],
- img_prefix=[data_root + 'VOC2007/', data_root + 'VOC2012/'],
- pipeline=train_pipeline)),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'VOC2007/ImageSets/Main/test.txt',
- img_prefix=data_root + 'VOC2007/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'VOC2007/ImageSets/Main/test.txt',
- img_prefix=data_root + 'VOC2007/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='mAP')
-# optimizer
-optimizer = dict(type='SGD', lr=1e-3, momentum=0.9, weight_decay=5e-4)
-optimizer_config = dict()
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 20])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/ssd300_voc'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/pascal_voc/ssd300_voc0712.py b/configs/pascal_voc/ssd300_voc0712.py
new file mode 100644
index 0000000000000000000000000000000000000000..5e59091ea638d1059fc0544fc7748be20b3f3f80
--- /dev/null
+++ b/configs/pascal_voc/ssd300_voc0712.py
@@ -0,0 +1,67 @@
+_base_ = [
+ '../_base_/models/ssd300.py', '../_base_/datasets/voc0712.py',
+ '../_base_/default_runtime.py'
+]
+model = dict(bbox_head=dict(num_classes=21, basesize_ratio_range=(0.2, 0.9)))
+# dataset settings
+dataset_type = 'VOCDataset'
+data_root = 'data/VOCdevkit/'
+img_norm_cfg = dict(mean=[123.675, 116.28, 103.53], std=[1, 1, 1], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile', to_float32=True),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(
+ type='PhotoMetricDistortion',
+ brightness_delta=32,
+ contrast_range=(0.5, 1.5),
+ saturation_range=(0.5, 1.5),
+ hue_delta=18),
+ dict(
+ type='Expand',
+ mean=img_norm_cfg['mean'],
+ to_rgb=img_norm_cfg['to_rgb'],
+ ratio_range=(1, 4)),
+ dict(
+ type='MinIoURandomCrop',
+ min_ious=(0.1, 0.3, 0.5, 0.7, 0.9),
+ min_crop_size=0.3),
+ dict(type='Resize', img_scale=(300, 300), keep_ratio=False),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(300, 300),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=False),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ imgs_per_gpu=8,
+ workers_per_gpu=3,
+ train=dict(
+ type='RepeatDataset', times=10, dataset=dict(pipeline=train_pipeline)),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+# optimizer
+optimizer = dict(type='SGD', lr=1e-3, momentum=0.9, weight_decay=5e-4)
+optimizer_config = dict()
+# learning policy
+lr_config = dict(
+ policy='step',
+ warmup='linear',
+ warmup_iters=500,
+ warmup_ratio=1.0 / 3,
+ step=[16, 20])
+checkpoint_config = dict(interval=1)
+# runtime settings
+total_epochs = 24
+work_dir = './work_dirs/ssd300_voc'
diff --git a/configs/pascal_voc/ssd512_voc.py b/configs/pascal_voc/ssd512_voc.py
deleted file mode 100644
index c81e5473ad82b343fe09df3ccd95771a75597027..0000000000000000000000000000000000000000
--- a/configs/pascal_voc/ssd512_voc.py
+++ /dev/null
@@ -1,138 +0,0 @@
-# model settings
-input_size = 512
-model = dict(
- type='SingleStageDetector',
- pretrained='open-mmlab://vgg16_caffe',
- backbone=dict(
- type='SSDVGG',
- input_size=input_size,
- depth=16,
- with_last_pool=False,
- ceil_mode=True,
- out_indices=(3, 4),
- out_feature_indices=(22, 34),
- l2_norm_scale=20),
- neck=None,
- bbox_head=dict(
- type='SSDHead',
- input_size=input_size,
- in_channels=(512, 1024, 512, 256, 256, 256, 256),
- num_classes=21,
- anchor_strides=(8, 16, 32, 64, 128, 256, 512),
- basesize_ratio_range=(0.15, 0.9),
- anchor_ratios=([2], [2, 3], [2, 3], [2, 3], [2, 3], [2], [2]),
- target_means=(.0, .0, .0, .0),
- target_stds=(0.1, 0.1, 0.2, 0.2)))
-# model training and testing settings
-cudnn_benchmark = True
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.,
- ignore_iof_thr=-1,
- gt_max_assign_all=False),
- smoothl1_beta=1.,
- allowed_border=-1,
- pos_weight=-1,
- neg_pos_ratio=3,
- debug=False)
-test_cfg = dict(
- nms=dict(type='nms', iou_thr=0.45),
- min_bbox_size=0,
- score_thr=0.02,
- max_per_img=200)
-# dataset settings
-dataset_type = 'VOCDataset'
-data_root = 'data/VOCdevkit/'
-img_norm_cfg = dict(mean=[123.675, 116.28, 103.53], std=[1, 1, 1], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile', to_float32=True),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(
- type='PhotoMetricDistortion',
- brightness_delta=32,
- contrast_range=(0.5, 1.5),
- saturation_range=(0.5, 1.5),
- hue_delta=18),
- dict(
- type='Expand',
- mean=img_norm_cfg['mean'],
- to_rgb=img_norm_cfg['to_rgb'],
- ratio_range=(1, 4)),
- dict(
- type='MinIoURandomCrop',
- min_ious=(0.1, 0.3, 0.5, 0.7, 0.9),
- min_crop_size=0.3),
- dict(type='Resize', img_scale=(512, 512), keep_ratio=False),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(512, 512),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=False),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=8,
- workers_per_gpu=3,
- train=dict(
- type='RepeatDataset',
- times=10,
- dataset=dict(
- type=dataset_type,
- ann_file=[
- data_root + 'VOC2007/ImageSets/Main/trainval.txt',
- data_root + 'VOC2012/ImageSets/Main/trainval.txt'
- ],
- img_prefix=[data_root + 'VOC2007/', data_root + 'VOC2012/'],
- pipeline=train_pipeline)),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'VOC2007/ImageSets/Main/test.txt',
- img_prefix=data_root + 'VOC2007/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'VOC2007/ImageSets/Main/test.txt',
- img_prefix=data_root + 'VOC2007/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='mAP')
-# optimizer
-optimizer = dict(type='SGD', lr=1e-3, momentum=0.9, weight_decay=5e-4)
-optimizer_config = dict()
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 20])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/ssd512_voc'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/pascal_voc/ssd512_voc0712.py b/configs/pascal_voc/ssd512_voc0712.py
new file mode 100644
index 0000000000000000000000000000000000000000..a308e00e0e5c50a9b7777dd2a63675db22c748ed
--- /dev/null
+++ b/configs/pascal_voc/ssd512_voc0712.py
@@ -0,0 +1,53 @@
+_base_ = 'ssd300_voc0712.py'
+input_size = 512
+model = dict(
+ backbone=dict(input_size=input_size),
+ bbox_head=dict(
+ input_size=input_size,
+ in_channels=(512, 1024, 512, 256, 256, 256, 256),
+ anchor_strides=(8, 16, 32, 64, 128, 256, 512),
+ basesize_ratio_range=(0.15, 0.9),
+ anchor_ratios=([2], [2, 3], [2, 3], [2, 3], [2, 3], [2], [2])))
+img_norm_cfg = dict(mean=[123.675, 116.28, 103.53], std=[1, 1, 1], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile', to_float32=True),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(
+ type='PhotoMetricDistortion',
+ brightness_delta=32,
+ contrast_range=(0.5, 1.5),
+ saturation_range=(0.5, 1.5),
+ hue_delta=18),
+ dict(
+ type='Expand',
+ mean=img_norm_cfg['mean'],
+ to_rgb=img_norm_cfg['to_rgb'],
+ ratio_range=(1, 4)),
+ dict(
+ type='MinIoURandomCrop',
+ min_ious=(0.1, 0.3, 0.5, 0.7, 0.9),
+ min_crop_size=0.3),
+ dict(type='Resize', img_scale=(512, 512), keep_ratio=False),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(512, 512),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=False),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ train=dict(dataset=dict(pipeline=train_pipeline)),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+work_dir = './work_dirs/ssd512_voc'
diff --git a/configs/reppoints/bbox_r50_grid_center_fpn_1x.py b/configs/reppoints/bbox_r50_grid_center_fpn_1x.py
deleted file mode 100644
index 8a94bb9f1b0c302005a38f5f6c5f440b54bf5c30..0000000000000000000000000000000000000000
--- a/configs/reppoints/bbox_r50_grid_center_fpn_1x.py
+++ /dev/null
@@ -1,143 +0,0 @@
-# model settings
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-
-model = dict(
- type='RepPointsDetector',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5,
- norm_cfg=norm_cfg),
- bbox_head=dict(
- type='RepPointsHead',
- num_classes=81,
- in_channels=256,
- feat_channels=256,
- point_feat_channels=256,
- stacked_convs=3,
- num_points=9,
- gradient_mul=0.1,
- point_strides=[8, 16, 32, 64, 128],
- point_base_scale=4,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox_init=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.5),
- loss_bbox_refine=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0),
- transform_method='minmax',
- use_grid_points=True))
-# training and testing settings
-train_cfg = dict(
- init=dict(
- assigner=dict(type='PointAssigner', scale=4, pos_num=1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False),
- refine=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/bbox_r50_grid_center_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/reppoints/bbox_r50_grid_center_fpn_gn-neck+head_1x_coco.py b/configs/reppoints/bbox_r50_grid_center_fpn_gn-neck+head_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..7bdcb8a2437790aa3c988122454c0c1b22ad2eea
--- /dev/null
+++ b/configs/reppoints/bbox_r50_grid_center_fpn_gn-neck+head_1x_coco.py
@@ -0,0 +1,3 @@
+_base_ = './reppoints_moment_r50_fpn_gn-neck+head_1x_coco.py'
+model = dict(bbox_head=dict(transform_method='minmax', use_grid_points=True))
+work_dir = './work_dirs/bbox_r50_grid_center_fpn_1x'
diff --git a/configs/reppoints/bbox_r50_grid_fpn_1x.py b/configs/reppoints/bbox_r50_grid_fpn_1x.py
deleted file mode 100644
index 5fe186a32fb14f101ce9d8a000c0d87e5cd4166e..0000000000000000000000000000000000000000
--- a/configs/reppoints/bbox_r50_grid_fpn_1x.py
+++ /dev/null
@@ -1,148 +0,0 @@
-# model settings
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-
-model = dict(
- type='RepPointsDetector',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5,
- norm_cfg=norm_cfg),
- bbox_head=dict(
- type='RepPointsHead',
- num_classes=81,
- in_channels=256,
- feat_channels=256,
- point_feat_channels=256,
- stacked_convs=3,
- num_points=9,
- gradient_mul=0.1,
- point_strides=[8, 16, 32, 64, 128],
- point_base_scale=4,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox_init=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.5),
- loss_bbox_refine=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0),
- transform_method='minmax',
- use_grid_points=True))
-# training and testing settings
-train_cfg = dict(
- init=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False),
- refine=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/bbox_r50_grid_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/reppoints/bbox_r50_grid_fpn_gn-neck+head_1x_coco.py b/configs/reppoints/bbox_r50_grid_fpn_gn-neck+head_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..96bcfecb4932c125c2110f6f8fcfbea967165170
--- /dev/null
+++ b/configs/reppoints/bbox_r50_grid_fpn_gn-neck+head_1x_coco.py
@@ -0,0 +1,13 @@
+_base_ = './reppoints_moment_r50_fpn_gn-neck+head_1x_coco.py'
+model = dict(bbox_head=dict(transform_method='minmax', use_grid_points=True))
+# training and testing settings
+train_cfg = dict(
+ init=dict(
+ assigner=dict(
+ _delete_=True,
+ type='MaxIoUAssigner',
+ pos_iou_thr=0.5,
+ neg_iou_thr=0.4,
+ min_pos_iou=0,
+ ignore_iof_thr=-1)))
+work_dir = './work_dirs/bbox_r50_grid_fpn_1x'
diff --git a/configs/reppoints/reppoints_minmax_r50_fpn_1x.py b/configs/reppoints/reppoints_minmax_r50_fpn_1x.py
deleted file mode 100644
index 974b749ed7ea537393edfc07598d92de263e04d2..0000000000000000000000000000000000000000
--- a/configs/reppoints/reppoints_minmax_r50_fpn_1x.py
+++ /dev/null
@@ -1,142 +0,0 @@
-# model settings
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-
-model = dict(
- type='RepPointsDetector',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5,
- norm_cfg=norm_cfg),
- bbox_head=dict(
- type='RepPointsHead',
- num_classes=81,
- in_channels=256,
- feat_channels=256,
- point_feat_channels=256,
- stacked_convs=3,
- num_points=9,
- gradient_mul=0.1,
- point_strides=[8, 16, 32, 64, 128],
- point_base_scale=4,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox_init=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.5),
- loss_bbox_refine=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0),
- transform_method='minmax'))
-# training and testing settings
-train_cfg = dict(
- init=dict(
- assigner=dict(type='PointAssigner', scale=4, pos_num=1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False),
- refine=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/reppoints_minmax_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/reppoints/reppoints_minmax_r50_fpn_gn-neck+head_1x_coco.py b/configs/reppoints/reppoints_minmax_r50_fpn_gn-neck+head_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..4e3df686840c9ef2fbf92a3d298c4d9e5752d6fb
--- /dev/null
+++ b/configs/reppoints/reppoints_minmax_r50_fpn_gn-neck+head_1x_coco.py
@@ -0,0 +1,3 @@
+_base_ = './reppoints_moment_r50_fpn_gn-neck+head_1x_coco.py'
+model = dict(bbox_head=dict(transform_method='minmax'))
+work_dir = './work_dirs/reppoints_minmax_r50_fpn_1x'
diff --git a/configs/reppoints/reppoints_moment_r101_dcn_fpn_2x.py b/configs/reppoints/reppoints_moment_r101_dcn_fpn_2x.py
deleted file mode 100644
index 44bf9fa576f194eef8be7c305e5380bee98e39ce..0000000000000000000000000000000000000000
--- a/configs/reppoints/reppoints_moment_r101_dcn_fpn_2x.py
+++ /dev/null
@@ -1,144 +0,0 @@
-# model settings
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-
-model = dict(
- type='RepPointsDetector',
- pretrained='torchvision://resnet101',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
- stage_with_dcn=(False, True, True, True)),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5,
- norm_cfg=norm_cfg),
- bbox_head=dict(
- type='RepPointsHead',
- num_classes=81,
- in_channels=256,
- feat_channels=256,
- point_feat_channels=256,
- stacked_convs=3,
- num_points=9,
- gradient_mul=0.1,
- point_strides=[8, 16, 32, 64, 128],
- point_base_scale=4,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox_init=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.5),
- loss_bbox_refine=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0),
- transform_method='moment'))
-# training and testing settings
-train_cfg = dict(
- init=dict(
- assigner=dict(type='PointAssigner', scale=4, pos_num=1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False),
- refine=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/reppoints_moment_r101_dcn_fpn_2x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/reppoints/reppoints_moment_r101_dcn_fpn_2x_mt.py b/configs/reppoints/reppoints_moment_r101_dcn_fpn_2x_mt.py
deleted file mode 100644
index 633d1b199389d75cde98d111b1b95f06f9d66710..0000000000000000000000000000000000000000
--- a/configs/reppoints/reppoints_moment_r101_dcn_fpn_2x_mt.py
+++ /dev/null
@@ -1,148 +0,0 @@
-# model settings
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-
-model = dict(
- type='RepPointsDetector',
- pretrained='torchvision://resnet101',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
- stage_with_dcn=(False, True, True, True)),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5,
- norm_cfg=norm_cfg),
- bbox_head=dict(
- type='RepPointsHead',
- num_classes=81,
- in_channels=256,
- feat_channels=256,
- point_feat_channels=256,
- stacked_convs=3,
- num_points=9,
- gradient_mul=0.1,
- point_strides=[8, 16, 32, 64, 128],
- point_base_scale=4,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox_init=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.5),
- loss_bbox_refine=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0),
- transform_method='moment'))
-# training and testing settings
-train_cfg = dict(
- init=dict(
- assigner=dict(type='PointAssigner', scale=4, pos_num=1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False),
- refine=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(
- type='Resize',
- img_scale=[(1333, 480), (1333, 960)],
- keep_ratio=True,
- multiscale_mode='range'),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/reppoints_moment_r101_dcn_fpn_2x_mt'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/reppoints/reppoints_moment_r101_fpn_2x.py b/configs/reppoints/reppoints_moment_r101_fpn_2x.py
deleted file mode 100644
index b610c3f21fb375061b4e2de54a0d099c53a7d3be..0000000000000000000000000000000000000000
--- a/configs/reppoints/reppoints_moment_r101_fpn_2x.py
+++ /dev/null
@@ -1,142 +0,0 @@
-# model settings
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-
-model = dict(
- type='RepPointsDetector',
- pretrained='torchvision://resnet101',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5,
- norm_cfg=norm_cfg),
- bbox_head=dict(
- type='RepPointsHead',
- num_classes=81,
- in_channels=256,
- feat_channels=256,
- point_feat_channels=256,
- stacked_convs=3,
- num_points=9,
- gradient_mul=0.1,
- point_strides=[8, 16, 32, 64, 128],
- point_base_scale=4,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox_init=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.5),
- loss_bbox_refine=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0),
- transform_method='moment'))
-# training and testing settings
-train_cfg = dict(
- init=dict(
- assigner=dict(type='PointAssigner', scale=4, pos_num=1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False),
- refine=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/reppoints_moment_r101_fpn_2x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/reppoints/reppoints_moment_r101_fpn_2x_mt.py b/configs/reppoints/reppoints_moment_r101_fpn_2x_mt.py
deleted file mode 100644
index 75dcc4606019d8909723b90c1df2ffd15e12f57a..0000000000000000000000000000000000000000
--- a/configs/reppoints/reppoints_moment_r101_fpn_2x_mt.py
+++ /dev/null
@@ -1,146 +0,0 @@
-# model settings
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-
-model = dict(
- type='RepPointsDetector',
- pretrained='torchvision://resnet101',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5,
- norm_cfg=norm_cfg),
- bbox_head=dict(
- type='RepPointsHead',
- num_classes=81,
- in_channels=256,
- feat_channels=256,
- point_feat_channels=256,
- stacked_convs=3,
- num_points=9,
- gradient_mul=0.1,
- point_strides=[8, 16, 32, 64, 128],
- point_base_scale=4,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox_init=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.5),
- loss_bbox_refine=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0),
- transform_method='moment'))
-# training and testing settings
-train_cfg = dict(
- init=dict(
- assigner=dict(type='PointAssigner', scale=4, pos_num=1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False),
- refine=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(
- type='Resize',
- img_scale=[(1333, 480), (1333, 960)],
- keep_ratio=True,
- multiscale_mode='range'),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/reppoints_moment_r101_fpn_2x_mt'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/reppoints/reppoints_moment_r101_fpn_dconv_c3-c5_gn-neck+head_2x_coco.py b/configs/reppoints/reppoints_moment_r101_fpn_dconv_c3-c5_gn-neck+head_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..e25d8dce8c97dd4e0f4262990e7adb0eeb2ece09
--- /dev/null
+++ b/configs/reppoints/reppoints_moment_r101_fpn_dconv_c3-c5_gn-neck+head_2x_coco.py
@@ -0,0 +1,8 @@
+_base_ = './reppoints_moment_r50_fpn_gn-neck+head_2x_coco.py'
+model = dict(
+ pretrained='torchvision://resnet101',
+ backbone=dict(
+ depth=101,
+ dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
+ stage_with_dcn=(False, True, True, True)))
+work_dir = './work_dirs/reppoints_moment_r101_dcn_fpn_2x'
diff --git a/configs/reppoints/reppoints_moment_r101_fpn_dconv_c3-c5_gn-neck+head_mstrain_480-960_2x_coco.py b/configs/reppoints/reppoints_moment_r101_fpn_dconv_c3-c5_gn-neck+head_mstrain_480-960_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..2511fd3e2b6ed3bd9262b327701f7f0cc363fe99
--- /dev/null
+++ b/configs/reppoints/reppoints_moment_r101_fpn_dconv_c3-c5_gn-neck+head_mstrain_480-960_2x_coco.py
@@ -0,0 +1,8 @@
+_base_ = './reppoints_moment_r50_fpn_gn-neck+head_mstrain_480-960_2x_coco.py'
+model = dict(
+ pretrained='torchvision://resnet101',
+ backbone=dict(
+ depth=101,
+ dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
+ stage_with_dcn=(False, True, True, True)))
+work_dir = './work_dirs/reppoints_moment_r101_dcn_fpn_2x_mt'
diff --git a/configs/reppoints/reppoints_moment_r101_fpn_gn-neck+head_2x_coco.py b/configs/reppoints/reppoints_moment_r101_fpn_gn-neck+head_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..7b0e031a085df11d7d414b975b8305620a56434c
--- /dev/null
+++ b/configs/reppoints/reppoints_moment_r101_fpn_gn-neck+head_2x_coco.py
@@ -0,0 +1,3 @@
+_base_ = './reppoints_moment_r50_fpn_gn-neck+head_2x_coco.py'
+model = dict(pretrained='torchvision://resnet101', backbone=dict(depth=101))
+work_dir = './work_dirs/reppoints_moment_r101_fpn_2x'
diff --git a/configs/reppoints/reppoints_moment_r101_fpn_gn-neck+head_mstrain_480-960_2x_coco.py b/configs/reppoints/reppoints_moment_r101_fpn_gn-neck+head_mstrain_480-960_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..fddd62da6a68eaeb7c9830fb5efbe841fcf76689
--- /dev/null
+++ b/configs/reppoints/reppoints_moment_r101_fpn_gn-neck+head_mstrain_480-960_2x_coco.py
@@ -0,0 +1,3 @@
+_base_ = './reppoints_moment_r50_fpn_gn-neck+head_mstrain_480-960_2x_coco.py'
+model = dict(pretrained='torchvision://resnet101', backbone=dict(depth=101))
+work_dir = './work_dirs/reppoints_moment_r101_fpn_2x_mt'
diff --git a/configs/reppoints/reppoints_moment_r50_fpn_1x.py b/configs/reppoints/reppoints_moment_r50_fpn_1x.py
deleted file mode 100644
index a2ce0284fc5c7ad141bc9f57fb706ea53628df64..0000000000000000000000000000000000000000
--- a/configs/reppoints/reppoints_moment_r50_fpn_1x.py
+++ /dev/null
@@ -1,142 +0,0 @@
-# model settings
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-
-model = dict(
- type='RepPointsDetector',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5,
- norm_cfg=norm_cfg),
- bbox_head=dict(
- type='RepPointsHead',
- num_classes=81,
- in_channels=256,
- feat_channels=256,
- point_feat_channels=256,
- stacked_convs=3,
- num_points=9,
- gradient_mul=0.1,
- point_strides=[8, 16, 32, 64, 128],
- point_base_scale=4,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox_init=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.5),
- loss_bbox_refine=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0),
- transform_method='moment'))
-# training and testing settings
-train_cfg = dict(
- init=dict(
- assigner=dict(type='PointAssigner', scale=4, pos_num=1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False),
- refine=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/reppoints_moment_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/reppoints/reppoints_moment_r50_fpn_1x_coco.py b/configs/reppoints/reppoints_moment_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..1e25c91ab7bb20dfb54be9012eb156208986fae2
--- /dev/null
+++ b/configs/reppoints/reppoints_moment_r50_fpn_1x_coco.py
@@ -0,0 +1,67 @@
+_base_ = [
+ '../_base_/datasets/coco_detection.py',
+ '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
+]
+model = dict(
+ type='RepPointsDetector',
+ pretrained='torchvision://resnet50',
+ backbone=dict(
+ type='ResNet',
+ depth=50,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'),
+ neck=dict(
+ type='FPN',
+ in_channels=[256, 512, 1024, 2048],
+ out_channels=256,
+ start_level=1,
+ add_extra_convs=True,
+ num_outs=5),
+ bbox_head=dict(
+ type='RepPointsHead',
+ num_classes=81,
+ in_channels=256,
+ feat_channels=256,
+ point_feat_channels=256,
+ stacked_convs=3,
+ num_points=9,
+ gradient_mul=0.1,
+ point_strides=[8, 16, 32, 64, 128],
+ point_base_scale=4,
+ loss_cls=dict(
+ type='FocalLoss',
+ use_sigmoid=True,
+ gamma=2.0,
+ alpha=0.25,
+ loss_weight=1.0),
+ loss_bbox_init=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.5),
+ loss_bbox_refine=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0),
+ transform_method='moment'))
+# training and testing settings
+train_cfg = dict(
+ init=dict(
+ assigner=dict(type='PointAssigner', scale=4, pos_num=1),
+ allowed_border=-1,
+ pos_weight=-1,
+ debug=False),
+ refine=dict(
+ assigner=dict(
+ type='MaxIoUAssigner',
+ pos_iou_thr=0.5,
+ neg_iou_thr=0.4,
+ min_pos_iou=0,
+ ignore_iof_thr=-1),
+ allowed_border=-1,
+ pos_weight=-1,
+ debug=False))
+test_cfg = dict(
+ nms_pre=1000,
+ min_bbox_size=0,
+ score_thr=0.05,
+ nms=dict(type='nms', iou_thr=0.5),
+ max_per_img=100)
+optimizer = dict(lr=0.01)
+work_dir = './work_dirs/reppoints_moment_r50_no_gn_fpn_1x'
diff --git a/configs/reppoints/reppoints_moment_r50_fpn_2x.py b/configs/reppoints/reppoints_moment_r50_fpn_2x.py
deleted file mode 100644
index cc51f3afe2b8fe66fc0878f6ab741b4fdfe6ba5d..0000000000000000000000000000000000000000
--- a/configs/reppoints/reppoints_moment_r50_fpn_2x.py
+++ /dev/null
@@ -1,142 +0,0 @@
-# model settings
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-
-model = dict(
- type='RepPointsDetector',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5,
- norm_cfg=norm_cfg),
- bbox_head=dict(
- type='RepPointsHead',
- num_classes=81,
- in_channels=256,
- feat_channels=256,
- point_feat_channels=256,
- stacked_convs=3,
- num_points=9,
- gradient_mul=0.1,
- point_strides=[8, 16, 32, 64, 128],
- point_base_scale=4,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox_init=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.5),
- loss_bbox_refine=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0),
- transform_method='moment'))
-# training and testing settings
-train_cfg = dict(
- init=dict(
- assigner=dict(type='PointAssigner', scale=4, pos_num=1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False),
- refine=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/reppoints_moment_r50_fpn_2x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/reppoints/reppoints_moment_r50_fpn_2x_mt.py b/configs/reppoints/reppoints_moment_r50_fpn_2x_mt.py
deleted file mode 100644
index 62ad7b0d7566a4f9098f26cf56f2f3ea1abce346..0000000000000000000000000000000000000000
--- a/configs/reppoints/reppoints_moment_r50_fpn_2x_mt.py
+++ /dev/null
@@ -1,146 +0,0 @@
-# model settings
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-
-model = dict(
- type='RepPointsDetector',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5,
- norm_cfg=norm_cfg),
- bbox_head=dict(
- type='RepPointsHead',
- num_classes=81,
- in_channels=256,
- feat_channels=256,
- point_feat_channels=256,
- stacked_convs=3,
- num_points=9,
- gradient_mul=0.1,
- point_strides=[8, 16, 32, 64, 128],
- point_base_scale=4,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox_init=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.5),
- loss_bbox_refine=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0),
- transform_method='moment'))
-# training and testing settings
-train_cfg = dict(
- init=dict(
- assigner=dict(type='PointAssigner', scale=4, pos_num=1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False),
- refine=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(
- type='Resize',
- img_scale=[(1333, 480), (1333, 960)],
- keep_ratio=True,
- multiscale_mode='range'),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/reppoints_moment_r50_fpn_2x_mt'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/reppoints/reppoints_moment_r50_fpn_gn-neck+head_1x_coco.py b/configs/reppoints/reppoints_moment_r50_fpn_gn-neck+head_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..fb41fe519faead7839bc3496da486eef55a7881b
--- /dev/null
+++ b/configs/reppoints/reppoints_moment_r50_fpn_gn-neck+head_1x_coco.py
@@ -0,0 +1,5 @@
+_base_ = './reppoints_moment_r50_fpn_1x_coco.py'
+norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
+model = dict(neck=dict(norm_cfg=norm_cfg), bbox_head=dict(norm_cfg=norm_cfg))
+optimizer = dict(lr=0.01)
+work_dir = './work_dirs/reppoints_moment_r50_fpn_1x'
diff --git a/configs/reppoints/reppoints_moment_r50_fpn_gn-neck+head_2x_coco.py b/configs/reppoints/reppoints_moment_r50_fpn_gn-neck+head_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..d5bdbe7ffdb13ab6eee9b84c257d8307c850b660
--- /dev/null
+++ b/configs/reppoints/reppoints_moment_r50_fpn_gn-neck+head_2x_coco.py
@@ -0,0 +1,4 @@
+_base_ = './reppoints_moment_r50_fpn_gn-neck+head_1x_coco.py'
+lr_config = dict(step=[16, 22])
+total_epochs = 24
+work_dir = './work_dirs/reppoints_moment_r50_fpn_2x'
diff --git a/configs/reppoints/reppoints_moment_r50_fpn_gn-neck+head_mstrain_480-960_2x_coco.py b/configs/reppoints/reppoints_moment_r50_fpn_gn-neck+head_mstrain_480-960_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..c72ff9acbe361881a0962629cbe3fb40312f7192
--- /dev/null
+++ b/configs/reppoints/reppoints_moment_r50_fpn_gn-neck+head_mstrain_480-960_2x_coco.py
@@ -0,0 +1,19 @@
+_base_ = './reppoints_moment_r50_fpn_gn-neck+head_2x_coco.py'
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(
+ type='Resize',
+ img_scale=[(1333, 480), (1333, 960)],
+ keep_ratio=True,
+ multiscale_mode='range'),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+data = dict(train=dict(pipeline=train_pipeline))
+work_dir = './work_dirs/reppoints_moment_r50_fpn_2x_mt'
diff --git a/configs/reppoints/reppoints_moment_r50_no_gn_fpn_1x.py b/configs/reppoints/reppoints_moment_r50_no_gn_fpn_1x.py
deleted file mode 100644
index bdaf07eefcb40aa762a98ed8aa49a49729b48ea1..0000000000000000000000000000000000000000
--- a/configs/reppoints/reppoints_moment_r50_no_gn_fpn_1x.py
+++ /dev/null
@@ -1,138 +0,0 @@
-# model settings
-model = dict(
- type='RepPointsDetector',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5),
- bbox_head=dict(
- type='RepPointsHead',
- num_classes=81,
- in_channels=256,
- feat_channels=256,
- point_feat_channels=256,
- stacked_convs=3,
- num_points=9,
- gradient_mul=0.1,
- point_strides=[8, 16, 32, 64, 128],
- point_base_scale=4,
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox_init=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.5),
- loss_bbox_refine=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0),
- transform_method='moment'))
-# training and testing settings
-train_cfg = dict(
- init=dict(
- assigner=dict(type='PointAssigner', scale=4, pos_num=1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False),
- refine=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/reppoints_moment_r50_no_gn_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/reppoints/reppoints_moment_x101_dcn_fpn_2x.py b/configs/reppoints/reppoints_moment_x101_dcn_fpn_2x.py
deleted file mode 100644
index acc23969c90f175521f43a5c336acd407f5dbeef..0000000000000000000000000000000000000000
--- a/configs/reppoints/reppoints_moment_x101_dcn_fpn_2x.py
+++ /dev/null
@@ -1,146 +0,0 @@
-# model settings
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-
-model = dict(
- type='RepPointsDetector',
- pretrained='open-mmlab://resnext101_32x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=32,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
- stage_with_dcn=(False, True, True, True)),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5,
- norm_cfg=norm_cfg),
- bbox_head=dict(
- type='RepPointsHead',
- num_classes=81,
- in_channels=256,
- feat_channels=256,
- point_feat_channels=256,
- stacked_convs=3,
- num_points=9,
- gradient_mul=0.1,
- point_strides=[8, 16, 32, 64, 128],
- point_base_scale=4,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox_init=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.5),
- loss_bbox_refine=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0),
- transform_method='moment'))
-# training and testing settings
-train_cfg = dict(
- init=dict(
- assigner=dict(type='PointAssigner', scale=4, pos_num=1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False),
- refine=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/reppoints_moment_x101_dcn_fpn_2x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/reppoints/reppoints_moment_x101_dcn_fpn_2x_mt.py b/configs/reppoints/reppoints_moment_x101_dcn_fpn_2x_mt.py
deleted file mode 100644
index 8fa745366befaca9e32185d43c2af0c25dd69348..0000000000000000000000000000000000000000
--- a/configs/reppoints/reppoints_moment_x101_dcn_fpn_2x_mt.py
+++ /dev/null
@@ -1,150 +0,0 @@
-# model settings
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-
-model = dict(
- type='RepPointsDetector',
- pretrained='open-mmlab://resnext101_32x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=32,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch',
- dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
- stage_with_dcn=(False, True, True, True)),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5,
- norm_cfg=norm_cfg),
- bbox_head=dict(
- type='RepPointsHead',
- num_classes=81,
- in_channels=256,
- feat_channels=256,
- point_feat_channels=256,
- stacked_convs=3,
- num_points=9,
- gradient_mul=0.1,
- point_strides=[8, 16, 32, 64, 128],
- point_base_scale=4,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox_init=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.5),
- loss_bbox_refine=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0),
- transform_method='moment'))
-# training and testing settings
-train_cfg = dict(
- init=dict(
- assigner=dict(type='PointAssigner', scale=4, pos_num=1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False),
- refine=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(
- type='Resize',
- img_scale=[(1333, 480), (1333, 960)],
- keep_ratio=True,
- multiscale_mode='range'),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/reppoints_moment_x101_dcn_fpn_2x_mt'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/reppoints/reppoints_moment_x101_fpn_dconv_c3-c5_gn-neck+head_2x_coco.py b/configs/reppoints/reppoints_moment_x101_fpn_dconv_c3-c5_gn-neck+head_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..707b8f979e60aeb0466f3418270e40839361a732
--- /dev/null
+++ b/configs/reppoints/reppoints_moment_x101_fpn_dconv_c3-c5_gn-neck+head_2x_coco.py
@@ -0,0 +1,16 @@
+_base_ = './reppoints_moment_r50_fpn_gn-neck+head_2x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_32x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=32,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch',
+ dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
+ stage_with_dcn=(False, True, True, True)))
+work_dir = './work_dirs/reppoints_moment_x101_dcn_fpn_2x'
diff --git a/configs/reppoints/reppoints_moment_x101_fpn_dconv_c3-c5_gn-neck+head_mstrain_480-960_2x_coco.py b/configs/reppoints/reppoints_moment_x101_fpn_dconv_c3-c5_gn-neck+head_mstrain_480-960_2x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..82521489171668a58ebc5b0c39b5d4d202668b31
--- /dev/null
+++ b/configs/reppoints/reppoints_moment_x101_fpn_dconv_c3-c5_gn-neck+head_mstrain_480-960_2x_coco.py
@@ -0,0 +1,16 @@
+_base_ = './reppoints_moment_r50_fpn_gn-neck+head_mstrain_480-960_2x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_32x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=32,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch',
+ dcn=dict(type='DCN', deformable_groups=1, fallback_on_stride=False),
+ stage_with_dcn=(False, True, True, True)))
+work_dir = './work_dirs/reppoints_moment_x101_dcn_fpn_2x_mt'
diff --git a/configs/reppoints/reppoints_partial_minmax_r50_fpn_1x.py b/configs/reppoints/reppoints_partial_minmax_r50_fpn_1x.py
deleted file mode 100644
index 659706133873ad9f17cfcbd8cde8d94ee91ec3d1..0000000000000000000000000000000000000000
--- a/configs/reppoints/reppoints_partial_minmax_r50_fpn_1x.py
+++ /dev/null
@@ -1,142 +0,0 @@
-# model settings
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-
-model = dict(
- type='RepPointsDetector',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5,
- norm_cfg=norm_cfg),
- bbox_head=dict(
- type='RepPointsHead',
- num_classes=81,
- in_channels=256,
- feat_channels=256,
- point_feat_channels=256,
- stacked_convs=3,
- num_points=9,
- gradient_mul=0.1,
- point_strides=[8, 16, 32, 64, 128],
- point_base_scale=4,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox_init=dict(type='SmoothL1Loss', beta=0.11, loss_weight=0.5),
- loss_bbox_refine=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0),
- transform_method='partial_minmax'))
-# training and testing settings
-train_cfg = dict(
- init=dict(
- assigner=dict(type='PointAssigner', scale=4, pos_num=1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False),
- refine=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/reppoints_partial_minmax_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/reppoints/reppoints_partial_minmax_r50_fpn_gn-neck+head_1x_coco.py b/configs/reppoints/reppoints_partial_minmax_r50_fpn_gn-neck+head_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..b09c7b8a6af303d908fa456eae2d1eaf4122daa3
--- /dev/null
+++ b/configs/reppoints/reppoints_partial_minmax_r50_fpn_gn-neck+head_1x_coco.py
@@ -0,0 +1,3 @@
+_base_ = './reppoints_moment_r50_fpn_gn-neck+head_1x_coco.py'
+model = dict(bbox_head=dict(transform_method='partial_minmax'))
+work_dir = './work_dirs/reppoints_partial_minmax_r50_fpn_1x'
diff --git a/configs/retinanet/retinanet_r101_fpn_1x_coco.py b/configs/retinanet/retinanet_r101_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..bcce4d8883e7c8399554a1a9b3e24ed003ff1cc4
--- /dev/null
+++ b/configs/retinanet/retinanet_r101_fpn_1x_coco.py
@@ -0,0 +1,3 @@
+_base_ = './retinanet_r50_fpn_1x_coco.py'
+model = dict(pretrained='torchvision://resnet101', backbone=dict(depth=101))
+work_dir = './work_dirs/retinanet_r101_fpn_1x'
diff --git a/configs/retinanet/retinanet_r50_caffe_fpn_1x_coco.py b/configs/retinanet/retinanet_r50_caffe_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..e56d2ec6abbd9ccda2ef3aba9aa8564c39b347ea
--- /dev/null
+++ b/configs/retinanet/retinanet_r50_caffe_fpn_1x_coco.py
@@ -0,0 +1,37 @@
+_base_ = './retinanet_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnet50_caffe',
+ backbone=dict(norm_cfg=dict(requires_grad=False), style='caffe'))
+# use caffe img_norm
+img_norm_cfg = dict(
+ mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(1333, 800),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ train=dict(pipeline=train_pipeline),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+work_dir = './work_dirs/retinanet_r50_fpn_1x'
diff --git a/configs/retinanet/retinanet_r50_fpn_1x_coco.py b/configs/retinanet/retinanet_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..88166c227f9780e19935cc0745f55ae618744ee1
--- /dev/null
+++ b/configs/retinanet/retinanet_r50_fpn_1x_coco.py
@@ -0,0 +1,8 @@
+_base_ = [
+ '../_base_/models/retinanet_r50_fpn.py',
+ '../_base_/datasets/coco_detection.py',
+ '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
+]
+# optimizer
+optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
+work_dir = './work_dirs/retinanet_r50_fpn_1x'
diff --git a/configs/retinanet/retinanet_x101_32x4d_fpn_1x_coco.py b/configs/retinanet/retinanet_x101_32x4d_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..aad5f119426783603e2dc1680b5b12ce5fe8bb86
--- /dev/null
+++ b/configs/retinanet/retinanet_x101_32x4d_fpn_1x_coco.py
@@ -0,0 +1,14 @@
+_base_ = './retinanet_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_32x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=32,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+work_dir = './work_dirs/retinanet_x101_32x4d_fpn_1x'
diff --git a/configs/retinanet/retinanet_x101_64x4d_fpn_1x_coco.py b/configs/retinanet/retinanet_x101_64x4d_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..31eb21ff3fe9811c8644b007c5a4bdef8c05ef4c
--- /dev/null
+++ b/configs/retinanet/retinanet_x101_64x4d_fpn_1x_coco.py
@@ -0,0 +1,14 @@
+_base_ = './retinanet_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_64x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=64,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+work_dir = './work_dirs/retinanet_x101_64x4d_fpn_1x'
diff --git a/configs/retinanet_r101_fpn_1x.py b/configs/retinanet_r101_fpn_1x.py
deleted file mode 100644
index 7e4e9919ab975d308266b6ccfbf1aa03d68c40c5..0000000000000000000000000000000000000000
--- a/configs/retinanet_r101_fpn_1x.py
+++ /dev/null
@@ -1,131 +0,0 @@
-# model settings
-model = dict(
- type='RetinaNet',
- pretrained='torchvision://resnet101',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5),
- bbox_head=dict(
- type='RetinaHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- octave_base_scale=4,
- scales_per_octave=3,
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[8, 16, 32, 64, 128],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/retinanet_r101_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/retinanet_r50_fpn_1x.py b/configs/retinanet_r50_fpn_1x.py
deleted file mode 100644
index 37f851ca0a7115cad5561272587274f0ca618e6d..0000000000000000000000000000000000000000
--- a/configs/retinanet_r50_fpn_1x.py
+++ /dev/null
@@ -1,131 +0,0 @@
-# model settings
-model = dict(
- type='RetinaNet',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5),
- bbox_head=dict(
- type='RetinaHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- octave_base_scale=4,
- scales_per_octave=3,
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[8, 16, 32, 64, 128],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/retinanet_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/retinanet_x101_32x4d_fpn_1x.py b/configs/retinanet_x101_32x4d_fpn_1x.py
deleted file mode 100644
index 3120f1c124da0e6a16c7931270d7e191e4f0e73e..0000000000000000000000000000000000000000
--- a/configs/retinanet_x101_32x4d_fpn_1x.py
+++ /dev/null
@@ -1,133 +0,0 @@
-# model settings
-model = dict(
- type='RetinaNet',
- pretrained='open-mmlab://resnext101_32x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=32,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5),
- bbox_head=dict(
- type='RetinaHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- octave_base_scale=4,
- scales_per_octave=3,
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[8, 16, 32, 64, 128],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/retinanet_x101_32x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/retinanet_x101_64x4d_fpn_1x.py b/configs/retinanet_x101_64x4d_fpn_1x.py
deleted file mode 100644
index 8c7e73da88ebc5e850d091ff1646c0bdab57f9d2..0000000000000000000000000000000000000000
--- a/configs/retinanet_x101_64x4d_fpn_1x.py
+++ /dev/null
@@ -1,133 +0,0 @@
-# model settings
-model = dict(
- type='RetinaNet',
- pretrained='open-mmlab://resnext101_64x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=64,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- start_level=1,
- add_extra_convs=True,
- num_outs=5),
- bbox_head=dict(
- type='RetinaHead',
- num_classes=81,
- in_channels=256,
- stacked_convs=4,
- feat_channels=256,
- octave_base_scale=4,
- scales_per_octave=3,
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[8, 16, 32, 64, 128],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='FocalLoss',
- use_sigmoid=True,
- gamma=2.0,
- alpha=0.25,
- loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=0.11, loss_weight=1.0)))
-# training and testing settings
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.4,
- min_pos_iou=0,
- ignore_iof_thr=-1),
- allowed_border=-1,
- pos_weight=-1,
- debug=False)
-test_cfg = dict(
- nms_pre=1000,
- min_bbox_size=0,
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100)
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/retinanet_x101_64x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/rpn/rpn_r101_fpn_1x_coco.py b/configs/rpn/rpn_r101_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..7e3223d97e9fb42d7383450bf92dfc4429c1b65b
--- /dev/null
+++ b/configs/rpn/rpn_r101_fpn_1x_coco.py
@@ -0,0 +1,3 @@
+_base_ = './rpn_r50_fpn_1x_coco.py'
+model = dict(pretrained='torchvision://resnet101', backbone=dict(depth=101))
+work_dir = './work_dirs/rpn_r101_fpn_1x'
diff --git a/configs/rpn/rpn_r50_caffe_c4_1x_coco.py b/configs/rpn/rpn_r50_caffe_c4_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..1697fed538a5f74cf4a98243b715e26ae8bbc5f9
--- /dev/null
+++ b/configs/rpn/rpn_r50_caffe_c4_1x_coco.py
@@ -0,0 +1,39 @@
+_base_ = [
+ '../_base_/models/rpn_r50_caffe_c4.py',
+ '../_base_/datasets/coco_detection.py',
+ '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
+]
+# dataset settings
+img_norm_cfg = dict(
+ mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True, with_label=False),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(1333, 800),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ train=dict(pipeline=train_pipeline),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+evaluation = dict(interval=1, metric='proposal_fast')
+work_dir = './work_dirs/rpn_r50_caffe_c4_1x'
diff --git a/configs/rpn/rpn_r50_caffe_fpn_1x_coco.py b/configs/rpn/rpn_r50_caffe_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..b07a42f453225ca192af22b8a86fcfc69a0af633
--- /dev/null
+++ b/configs/rpn/rpn_r50_caffe_fpn_1x_coco.py
@@ -0,0 +1,37 @@
+_base_ = './rpn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnet50_caffe',
+ backbone=dict(norm_cfg=dict(requires_grad=False), style='caffe'))
+# use caffe img_norm
+img_norm_cfg = dict(
+ mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True, with_label=False),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(1333, 800),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=True),
+ dict(type='RandomFlip'),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ train=dict(pipeline=train_pipeline),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+work_dir = './work_dirs/rpn_r50_caffe_fpn_1x'
diff --git a/configs/rpn/rpn_r50_fpn_1x_coco.py b/configs/rpn/rpn_r50_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..5e459d88d2188479e579a85047aebb48d4cfc829
--- /dev/null
+++ b/configs/rpn/rpn_r50_fpn_1x_coco.py
@@ -0,0 +1,19 @@
+_base_ = [
+ '../_base_/models/rpn_r50_fpn.py', '../_base_/datasets/coco_detection.py',
+ '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
+]
+img_norm_cfg = dict(
+ mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(type='LoadAnnotations', with_bbox=True, with_label=False),
+ dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='Pad', size_divisor=32),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes']),
+]
+data = dict(train=dict(pipeline=train_pipeline))
+evaluation = dict(interval=1, metric='proposal_fast')
+work_dir = './work_dirs/rpn_r50_fpn_1x'
diff --git a/configs/rpn/rpn_x101_32x4d_fpn_1x_coco.py b/configs/rpn/rpn_x101_32x4d_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..7dc2bd6a49d72b4e5f58139d4da58fe4e48cac98
--- /dev/null
+++ b/configs/rpn/rpn_x101_32x4d_fpn_1x_coco.py
@@ -0,0 +1,14 @@
+_base_ = './rpn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_32x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=32,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+work_dir = './work_dirs/rpn_x101_32x4d_fpn_1x'
diff --git a/configs/rpn/rpn_x101_64x4d_fpn_1x_coco.py b/configs/rpn/rpn_x101_64x4d_fpn_1x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..b25888c62ae6bfa758bc685e406c4cc63fa63e29
--- /dev/null
+++ b/configs/rpn/rpn_x101_64x4d_fpn_1x_coco.py
@@ -0,0 +1,14 @@
+_base_ = './rpn_r50_fpn_1x_coco.py'
+model = dict(
+ pretrained='open-mmlab://resnext101_64x4d',
+ backbone=dict(
+ type='ResNeXt',
+ depth=101,
+ groups=64,
+ base_width=4,
+ num_stages=4,
+ out_indices=(0, 1, 2, 3),
+ frozen_stages=1,
+ norm_cfg=dict(type='BN', requires_grad=True),
+ style='pytorch'))
+work_dir = './work_dirs/rpn_x101_64x4d_fpn_1x'
diff --git a/configs/rpn_r101_fpn_1x.py b/configs/rpn_r101_fpn_1x.py
deleted file mode 100644
index 498dff4eecfa578be458bb457cd4aa49bd77b2cb..0000000000000000000000000000000000000000
--- a/configs/rpn_r101_fpn_1x.py
+++ /dev/null
@@ -1,131 +0,0 @@
-# model settings
-model = dict(
- type='RPN',
- pretrained='torchvision://resnet101',
- backbone=dict(
- type='ResNet',
- depth=101,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_label=False),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='proposal_fast')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-# runner configs
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/rpn_r101_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/rpn_r50_caffe_c4_1x.py b/configs/rpn_r50_caffe_c4_1x.py
deleted file mode 100644
index c0b05a8a994d137131a6ac5abd4d1970786c3925..0000000000000000000000000000000000000000
--- a/configs/rpn_r50_caffe_c4_1x.py
+++ /dev/null
@@ -1,130 +0,0 @@
-# model settings
-model = dict(
- type='RPN',
- pretrained='open-mmlab://resnet50_caffe',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=3,
- strides=(1, 2, 2),
- dilations=(1, 1, 1),
- out_indices=(2, ),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=False),
- norm_eval=True,
- style='caffe'),
- neck=None,
- rpn_head=dict(
- type='RPNHead',
- in_channels=1024,
- feat_channels=1024,
- anchor_scales=[2, 4, 8, 16, 32],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[16],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=12000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[102.9801, 115.9465, 122.7717], std=[1.0, 1.0, 1.0], to_rgb=False)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_label=False),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='proposal_fast')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-# runner configs
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/rpn_r50_caffe_c4_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/rpn_r50_fpn_1x.py b/configs/rpn_r50_fpn_1x.py
deleted file mode 100644
index d0bedd6bc07f0fbc7732adf9ddff44fcfd0fcc9f..0000000000000000000000000000000000000000
--- a/configs/rpn_r50_fpn_1x.py
+++ /dev/null
@@ -1,131 +0,0 @@
-# model settings
-model = dict(
- type='RPN',
- pretrained='torchvision://resnet50',
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_label=False),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='proposal_fast')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-# runner configs
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/rpn_r50_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/rpn_x101_32x4d_fpn_1x.py b/configs/rpn_x101_32x4d_fpn_1x.py
deleted file mode 100644
index ab24b69f3fc33f1a466e3846f62ac256c0f3862d..0000000000000000000000000000000000000000
--- a/configs/rpn_x101_32x4d_fpn_1x.py
+++ /dev/null
@@ -1,133 +0,0 @@
-# model settings
-model = dict(
- type='RPN',
- pretrained='open-mmlab://resnext101_32x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=32,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_label=False),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='proposal_fast')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-# runner configs
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/rpn_x101_32x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/rpn_x101_64x4d_fpn_1x.py b/configs/rpn_x101_64x4d_fpn_1x.py
deleted file mode 100644
index 98d61531063aff5a8e20cdabb22853e08bb15f57..0000000000000000000000000000000000000000
--- a/configs/rpn_x101_64x4d_fpn_1x.py
+++ /dev/null
@@ -1,133 +0,0 @@
-# model settings
-model = dict(
- type='RPN',
- pretrained='open-mmlab://resnext101_64x4d',
- backbone=dict(
- type='ResNeXt',
- depth=101,
- groups=64,
- base_width=4,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=1,
- norm_cfg=dict(type='BN', requires_grad=True),
- style='pytorch'),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_label=False),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='proposal_fast')
-# optimizer
-optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
-# runner configs
-optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[8, 11])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 12
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/rpn_x101_64x4d_fpn_1x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/scratch/faster_rcnn_r50_fpn_gn-all_scratch_6x_coco.py b/configs/scratch/faster_rcnn_r50_fpn_gn-all_scratch_6x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..e939670ca3625e893d912bda091a75aa06b7db3a
--- /dev/null
+++ b/configs/scratch/faster_rcnn_r50_fpn_gn-all_scratch_6x_coco.py
@@ -0,0 +1,22 @@
+_base_ = [
+ '../_base_/models/faster_rcnn_r50_fpn.py',
+ '../_base_/datasets/coco_detection.py',
+ '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
+]
+norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
+model = dict(
+ pretrained=None,
+ backbone=dict(
+ frozen_stages=-1, zero_init_residual=False, norm_cfg=norm_cfg),
+ neck=dict(norm_cfg=norm_cfg),
+ bbox_head=dict(
+ type='Shared4Conv1FCBBoxHead',
+ conv_out_channels=256,
+ norm_cfg=norm_cfg))
+# optimizer
+optimizer = dict(paramwise_options=dict(norm_decay_mult=0))
+optimizer_config = dict(_delete_=True, grad_clip=None)
+# learning policy
+lr_config = dict(warmup_ratio=0.1, step=[65, 71])
+total_epochs = 73
+work_dir = './work_dirs/scratch_faster_rcnn_r50_fpn_gn_6x'
diff --git a/configs/scratch/mask_rcnn_r50_fpn_gn-all_scratch_6x_coco.py b/configs/scratch/mask_rcnn_r50_fpn_gn-all_scratch_6x_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..30f1ef02da168155bf8aaf2751935122e90dd9e9
--- /dev/null
+++ b/configs/scratch/mask_rcnn_r50_fpn_gn-all_scratch_6x_coco.py
@@ -0,0 +1,23 @@
+_base_ = [
+ '../_base_/models/mask_rcnn_r50_fpn.py',
+ '../_base_/datasets/coco_instance.py',
+ '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
+]
+norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
+model = dict(
+ pretrained=None,
+ backbone=dict(
+ frozen_stages=-1, zero_init_residual=False, norm_cfg=norm_cfg),
+ neck=dict(norm_cfg=norm_cfg),
+ bbox_head=dict(
+ type='Shared4Conv1FCBBoxHead',
+ conv_out_channels=256,
+ norm_cfg=norm_cfg),
+ mask_head=dict(norm_cfg=norm_cfg))
+# optimizer
+optimizer = dict(paramwise_options=dict(norm_decay_mult=0))
+optimizer_config = dict(_delete_=True, grad_clip=None)
+# learning policy
+lr_config = dict(warmup_ratio=0.1, step=[65, 71])
+total_epochs = 73
+work_dir = './work_dirs/scratch_mask_rcnn_r50_fpn_gn_6x'
diff --git a/configs/scratch/scratch_faster_rcnn_r50_fpn_gn_6x.py b/configs/scratch/scratch_faster_rcnn_r50_fpn_gn_6x.py
deleted file mode 100644
index 2a5935148d6d0531946bf8d783a0b23717484c99..0000000000000000000000000000000000000000
--- a/configs/scratch/scratch_faster_rcnn_r50_fpn_gn_6x.py
+++ /dev/null
@@ -1,185 +0,0 @@
-# model settings
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-model = dict(
- type='FasterRCNN',
- pretrained=None,
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=-1,
- style='pytorch',
- zero_init_residual=False,
- norm_cfg=norm_cfg),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5,
- norm_cfg=norm_cfg),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='ConvFCBBoxHead',
- num_shared_convs=4,
- num_shared_fcs=1,
- in_channels=256,
- conv_out_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(
- type='SGD',
- lr=0.02,
- momentum=0.9,
- weight_decay=0.0001,
- paramwise_options=dict(norm_decay_mult=0))
-optimizer_config = dict(grad_clip=None)
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=0.1,
- step=[65, 71])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 73
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/scratch_faster_rcnn_r50_fpn_gn_6x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/scratch/scratch_mask_rcnn_r50_fpn_gn_6x.py b/configs/scratch/scratch_mask_rcnn_r50_fpn_gn_6x.py
deleted file mode 100644
index 1f389cd58fd05fe35e83c344b63924bfb8014a85..0000000000000000000000000000000000000000
--- a/configs/scratch/scratch_mask_rcnn_r50_fpn_gn_6x.py
+++ /dev/null
@@ -1,202 +0,0 @@
-# model settings
-norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)
-model = dict(
- type='MaskRCNN',
- pretrained=None,
- backbone=dict(
- type='ResNet',
- depth=50,
- num_stages=4,
- out_indices=(0, 1, 2, 3),
- frozen_stages=-1,
- style='pytorch',
- zero_init_residual=False,
- norm_cfg=norm_cfg),
- neck=dict(
- type='FPN',
- in_channels=[256, 512, 1024, 2048],
- out_channels=256,
- num_outs=5,
- norm_cfg=norm_cfg),
- rpn_head=dict(
- type='RPNHead',
- in_channels=256,
- feat_channels=256,
- anchor_scales=[8],
- anchor_ratios=[0.5, 1.0, 2.0],
- anchor_strides=[4, 8, 16, 32, 64],
- target_means=[.0, .0, .0, .0],
- target_stds=[1.0, 1.0, 1.0, 1.0],
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
- bbox_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- bbox_head=dict(
- type='ConvFCBBoxHead',
- num_shared_convs=4,
- num_shared_fcs=1,
- in_channels=256,
- conv_out_channels=256,
- fc_out_channels=1024,
- roi_feat_size=7,
- num_classes=81,
- target_means=[0., 0., 0., 0.],
- target_stds=[0.1, 0.1, 0.2, 0.2],
- reg_class_agnostic=False,
- norm_cfg=norm_cfg,
- loss_cls=dict(
- type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
- loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)),
- mask_roi_extractor=dict(
- type='SingleRoIExtractor',
- roi_layer=dict(type='RoIAlign', out_size=14, sample_num=2),
- out_channels=256,
- featmap_strides=[4, 8, 16, 32]),
- mask_head=dict(
- type='FCNMaskHead',
- num_convs=4,
- in_channels=256,
- conv_out_channels=256,
- num_classes=81,
- norm_cfg=norm_cfg,
- loss_mask=dict(
- type='CrossEntropyLoss', use_mask=True, loss_weight=1.0)))
-# model training and testing settings
-train_cfg = dict(
- rpn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.7,
- neg_iou_thr=0.3,
- min_pos_iou=0.3,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=256,
- pos_fraction=0.5,
- neg_pos_ub=-1,
- add_gt_as_proposals=False),
- allowed_border=0,
- pos_weight=-1,
- debug=False),
- rpn_proposal=dict(
- nms_across_levels=False,
- nms_pre=2000,
- nms_post=2000,
- max_num=2000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.5,
- ignore_iof_thr=-1),
- sampler=dict(
- type='RandomSampler',
- num=512,
- pos_fraction=0.25,
- neg_pos_ub=-1,
- add_gt_as_proposals=True),
- mask_size=28,
- pos_weight=-1,
- debug=False))
-test_cfg = dict(
- rpn=dict(
- nms_across_levels=False,
- nms_pre=1000,
- nms_post=1000,
- max_num=1000,
- nms_thr=0.7,
- min_bbox_size=0),
- rcnn=dict(
- score_thr=0.05,
- nms=dict(type='nms', iou_thr=0.5),
- max_per_img=100,
- mask_thr_binary=0.5))
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(
- mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(type='LoadAnnotations', with_bbox=True, with_mask=True),
- dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(1333, 800),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=True),
- dict(type='RandomFlip'),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='Pad', size_divisor=32),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=2,
- workers_per_gpu=2,
- train=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric=['bbox', 'segm'])
-# optimizer
-optimizer = dict(
- type='SGD',
- lr=0.02,
- momentum=0.9,
- weight_decay=0.0001,
- paramwise_options=dict(norm_decay_mult=0))
-optimizer_config = dict(grad_clip=None)
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=0.1,
- step=[65, 71])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 73
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/scratch_mask_rcnn_r50_fpn_gn_6x'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/ssd/ssd300_coco.py b/configs/ssd/ssd300_coco.py
new file mode 100644
index 0000000000000000000000000000000000000000..3b367d3c02c5fb50b246d8ade4357e15f8120391
--- /dev/null
+++ b/configs/ssd/ssd300_coco.py
@@ -0,0 +1,63 @@
+_base_ = [
+ '../_base_/models/ssd300.py', '../_base_/datasets/coco_detection.py',
+ '../_base_/schedules/schedule_2x.py', '../_base_/default_runtime.py'
+]
+# dataset settings
+dataset_type = 'CocoDataset'
+data_root = 'data/coco/'
+img_norm_cfg = dict(mean=[123.675, 116.28, 103.53], std=[1, 1, 1], to_rgb=True)
+train_pipeline = [
+ dict(type='LoadImageFromFile', to_float32=True),
+ dict(type='LoadAnnotations', with_bbox=True),
+ dict(
+ type='PhotoMetricDistortion',
+ brightness_delta=32,
+ contrast_range=(0.5, 1.5),
+ saturation_range=(0.5, 1.5),
+ hue_delta=18),
+ dict(
+ type='Expand',
+ mean=img_norm_cfg['mean'],
+ to_rgb=img_norm_cfg['to_rgb'],
+ ratio_range=(1, 4)),
+ dict(
+ type='MinIoURandomCrop',
+ min_ious=(0.1, 0.3, 0.5, 0.7, 0.9),
+ min_crop_size=0.3),
+ dict(type='Resize', img_scale=(300, 300), keep_ratio=False),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='RandomFlip', flip_ratio=0.5),
+ dict(type='DefaultFormatBundle'),
+ dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+]
+test_pipeline = [
+ dict(type='LoadImageFromFile'),
+ dict(
+ type='MultiScaleFlipAug',
+ img_scale=(300, 300),
+ flip=False,
+ transforms=[
+ dict(type='Resize', keep_ratio=False),
+ dict(type='Normalize', **img_norm_cfg),
+ dict(type='ImageToTensor', keys=['img']),
+ dict(type='Collect', keys=['img']),
+ ])
+]
+data = dict(
+ imgs_per_gpu=8,
+ workers_per_gpu=3,
+ train=dict(
+ _delete_=True,
+ type='RepeatDataset',
+ times=5,
+ dataset=dict(
+ type=dataset_type,
+ ann_file=data_root + 'annotations/instances_train2017.json',
+ img_prefix=data_root + 'train2017/',
+ pipeline=train_pipeline)),
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
+# optimizer
+optimizer = dict(type='SGD', lr=2e-3, momentum=0.9, weight_decay=5e-4)
+optimizer_config = dict(_delete_=True)
+work_dir = './work_dirs/ssd300_coco'
diff --git a/configs/ssd512_coco.py b/configs/ssd/ssd512_coco.py
similarity index 52%
rename from configs/ssd512_coco.py
rename to configs/ssd/ssd512_coco.py
index e6534c516168fe7e9d91ff7f60436b2e283f84cc..13ea14f3f5576db5f9aea3a28ee7654fd6606de8 100644
--- a/configs/ssd512_coco.py
+++ b/configs/ssd/ssd512_coco.py
@@ -1,48 +1,13 @@
-# model settings
+_base_ = 'ssd300_coco.py'
input_size = 512
model = dict(
- type='SingleStageDetector',
- pretrained='open-mmlab://vgg16_caffe',
- backbone=dict(
- type='SSDVGG',
- input_size=input_size,
- depth=16,
- with_last_pool=False,
- ceil_mode=True,
- out_indices=(3, 4),
- out_feature_indices=(22, 34),
- l2_norm_scale=20),
- neck=None,
+ backbone=dict(input_size=input_size),
bbox_head=dict(
- type='SSDHead',
input_size=input_size,
in_channels=(512, 1024, 512, 256, 256, 256, 256),
- num_classes=81,
anchor_strides=(8, 16, 32, 64, 128, 256, 512),
basesize_ratio_range=(0.1, 0.9),
- anchor_ratios=([2], [2, 3], [2, 3], [2, 3], [2, 3], [2], [2]),
- target_means=(.0, .0, .0, .0),
- target_stds=(0.1, 0.1, 0.2, 0.2)))
-cudnn_benchmark = True
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.,
- ignore_iof_thr=-1,
- gt_max_assign_all=False),
- smoothl1_beta=1.,
- allowed_border=-1,
- pos_weight=-1,
- neg_pos_ratio=3,
- debug=False)
-test_cfg = dict(
- nms=dict(type='nms', iou_thr=0.45),
- min_bbox_size=0,
- score_thr=0.02,
- max_per_img=200)
-# model training and testing settings
+ anchor_ratios=([2], [2, 3], [2, 3], [2, 3], [2, 3], [2], [2])))
# dataset settings
dataset_type = 'CocoDataset'
data_root = 'data/coco/'
@@ -88,6 +53,7 @@ data = dict(
imgs_per_gpu=8,
workers_per_gpu=3,
train=dict(
+ _delete_=True,
type='RepeatDataset',
times=5,
dataset=dict(
@@ -95,41 +61,9 @@ data = dict(
ann_file=data_root + 'annotations/instances_train2017.json',
img_prefix=data_root + 'train2017/',
pipeline=train_pipeline)),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
+ val=dict(pipeline=test_pipeline),
+ test=dict(pipeline=test_pipeline))
# optimizer
optimizer = dict(type='SGD', lr=2e-3, momentum=0.9, weight_decay=5e-4)
-optimizer_config = dict()
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
+optimizer_config = dict(_delete_=True)
work_dir = './work_dirs/ssd512_coco'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/ssd300_coco.py b/configs/ssd300_coco.py
deleted file mode 100644
index 35219fac316e419149f78c6a0be27d34600e951a..0000000000000000000000000000000000000000
--- a/configs/ssd300_coco.py
+++ /dev/null
@@ -1,135 +0,0 @@
-# model settings
-input_size = 300
-model = dict(
- type='SingleStageDetector',
- pretrained='open-mmlab://vgg16_caffe',
- backbone=dict(
- type='SSDVGG',
- input_size=input_size,
- depth=16,
- with_last_pool=False,
- ceil_mode=True,
- out_indices=(3, 4),
- out_feature_indices=(22, 34),
- l2_norm_scale=20),
- neck=None,
- bbox_head=dict(
- type='SSDHead',
- input_size=input_size,
- in_channels=(512, 1024, 512, 256, 256, 256),
- num_classes=81,
- anchor_strides=(8, 16, 32, 64, 100, 300),
- basesize_ratio_range=(0.15, 0.9),
- anchor_ratios=([2], [2, 3], [2, 3], [2, 3], [2], [2]),
- target_means=(.0, .0, .0, .0),
- target_stds=(0.1, 0.1, 0.2, 0.2)))
-cudnn_benchmark = True
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.,
- ignore_iof_thr=-1,
- gt_max_assign_all=False),
- smoothl1_beta=1.,
- allowed_border=-1,
- pos_weight=-1,
- neg_pos_ratio=3,
- debug=False)
-test_cfg = dict(
- nms=dict(type='nms', iou_thr=0.45),
- min_bbox_size=0,
- score_thr=0.02,
- max_per_img=200)
-# model training and testing settings
-# dataset settings
-dataset_type = 'CocoDataset'
-data_root = 'data/coco/'
-img_norm_cfg = dict(mean=[123.675, 116.28, 103.53], std=[1, 1, 1], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile', to_float32=True),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(
- type='PhotoMetricDistortion',
- brightness_delta=32,
- contrast_range=(0.5, 1.5),
- saturation_range=(0.5, 1.5),
- hue_delta=18),
- dict(
- type='Expand',
- mean=img_norm_cfg['mean'],
- to_rgb=img_norm_cfg['to_rgb'],
- ratio_range=(1, 4)),
- dict(
- type='MinIoURandomCrop',
- min_ious=(0.1, 0.3, 0.5, 0.7, 0.9),
- min_crop_size=0.3),
- dict(type='Resize', img_scale=(300, 300), keep_ratio=False),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
-]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(300, 300),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=False),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=8,
- workers_per_gpu=3,
- train=dict(
- type='RepeatDataset',
- times=5,
- dataset=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_train2017.json',
- img_prefix=data_root + 'train2017/',
- pipeline=train_pipeline)),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'annotations/instances_val2017.json',
- img_prefix=data_root + 'val2017/',
- pipeline=test_pipeline))
-evaluation = dict(interval=1, metric='bbox')
-# optimizer
-optimizer = dict(type='SGD', lr=2e-3, momentum=0.9, weight_decay=5e-4)
-optimizer_config = dict()
-# learning policy
-lr_config = dict(
- policy='step',
- warmup='linear',
- warmup_iters=500,
- warmup_ratio=1.0 / 3,
- step=[16, 22])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=50,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
-# runtime settings
-total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
-work_dir = './work_dirs/ssd300_coco'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/configs/wider_face/ssd300_wider_face.py b/configs/wider_face/ssd300_wider_face.py
index 9c0ce79d0bc4213f551179094274cc83abb19767..3f65ec1adf64276a8333a92db32dd4ee4809c5df 100644
--- a/configs/wider_face/ssd300_wider_face.py
+++ b/configs/wider_face/ssd300_wider_face.py
@@ -1,111 +1,8 @@
-# model settings
-input_size = 300
-model = dict(
- type='SingleStageDetector',
- pretrained='open-mmlab://vgg16_caffe',
- backbone=dict(
- type='SSDVGG',
- input_size=input_size,
- depth=16,
- with_last_pool=False,
- ceil_mode=True,
- out_indices=(3, 4),
- out_feature_indices=(22, 34),
- l2_norm_scale=20),
- neck=None,
- bbox_head=dict(
- type='SSDHead',
- input_size=input_size,
- in_channels=(512, 1024, 512, 256, 256, 256),
- num_classes=2,
- anchor_strides=(8, 16, 32, 64, 100, 300),
- basesize_ratio_range=(0.15, 0.9),
- anchor_ratios=([2], [2, 3], [2, 3], [2, 3], [2], [2]),
- target_means=(.0, .0, .0, .0),
- target_stds=(0.1, 0.1, 0.2, 0.2)))
-# model training and testing settings
-cudnn_benchmark = True
-train_cfg = dict(
- assigner=dict(
- type='MaxIoUAssigner',
- pos_iou_thr=0.5,
- neg_iou_thr=0.5,
- min_pos_iou=0.,
- ignore_iof_thr=-1,
- gt_max_assign_all=False),
- smoothl1_beta=1.,
- allowed_border=-1,
- pos_weight=-1,
- neg_pos_ratio=3,
- debug=False)
-test_cfg = dict(
- nms=dict(type='nms', iou_thr=0.45),
- min_bbox_size=0,
- score_thr=0.02,
- max_per_img=200)
-# dataset settings
-dataset_type = 'WIDERFaceDataset'
-data_root = 'data/WIDERFace/'
-img_norm_cfg = dict(mean=[123.675, 116.28, 103.53], std=[1, 1, 1], to_rgb=True)
-train_pipeline = [
- dict(type='LoadImageFromFile', to_float32=True),
- dict(type='LoadAnnotations', with_bbox=True),
- dict(
- type='PhotoMetricDistortion',
- brightness_delta=32,
- contrast_range=(0.5, 1.5),
- saturation_range=(0.5, 1.5),
- hue_delta=18),
- dict(
- type='Expand',
- mean=img_norm_cfg['mean'],
- to_rgb=img_norm_cfg['to_rgb'],
- ratio_range=(1, 4)),
- dict(
- type='MinIoURandomCrop',
- min_ious=(0.1, 0.3, 0.5, 0.7, 0.9),
- min_crop_size=0.3),
- dict(type='Resize', img_scale=(300, 300), keep_ratio=False),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='RandomFlip', flip_ratio=0.5),
- dict(type='DefaultFormatBundle'),
- dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
+_base_ = [
+ '../_base_/models/ssd300.py', '../_base_/datasets/wider_face.py',
+ '../_base_/default_runtime.py'
]
-test_pipeline = [
- dict(type='LoadImageFromFile'),
- dict(
- type='MultiScaleFlipAug',
- img_scale=(300, 300),
- flip=False,
- transforms=[
- dict(type='Resize', keep_ratio=False),
- dict(type='Normalize', **img_norm_cfg),
- dict(type='ImageToTensor', keys=['img']),
- dict(type='Collect', keys=['img']),
- ])
-]
-data = dict(
- imgs_per_gpu=60,
- workers_per_gpu=2,
- train=dict(
- type='RepeatDataset',
- times=2,
- dataset=dict(
- type=dataset_type,
- ann_file=data_root + 'train.txt',
- img_prefix=data_root + 'WIDER_train/',
- min_size=17,
- pipeline=train_pipeline)),
- val=dict(
- type=dataset_type,
- ann_file=data_root + 'val.txt',
- img_prefix=data_root + 'WIDER_val/',
- pipeline=test_pipeline),
- test=dict(
- type=dataset_type,
- ann_file=data_root + 'val.txt',
- img_prefix=data_root + 'WIDER_val/',
- pipeline=test_pipeline))
+model = dict(bbox_head=dict(num_classes=2))
# optimizer
optimizer = dict(type='SGD', lr=0.012, momentum=0.9, weight_decay=5e-4)
optimizer_config = dict()
@@ -116,20 +13,7 @@ lr_config = dict(
warmup_iters=1000,
warmup_ratio=1.0 / 3,
step=[16, 20])
-checkpoint_config = dict(interval=1)
-# yapf:disable
-log_config = dict(
- interval=1,
- hooks=[
- dict(type='TextLoggerHook'),
- # dict(type='TensorboardLoggerHook')
- ])
-# yapf:enable
# runtime settings
total_epochs = 24
-dist_params = dict(backend='nccl')
-log_level = 'INFO'
+log_config = dict(interval=1)
work_dir = './work_dirs/ssd300_wider'
-load_from = None
-resume_from = None
-workflow = [('train', 1)]
diff --git a/demo/inference_demo.ipynb b/demo/inference_demo.ipynb
index ffe4765de5ffb037fcba89c20275c35fc08ffc77..fa6d513c5c0c4cc040b9639a96cb9e2ba89c1d51 100644
--- a/demo/inference_demo.ipynb
+++ b/demo/inference_demo.ipynb
@@ -16,7 +16,7 @@
"metadata": {},
"outputs": [],
"source": [
- "config_file = '../configs/faster_rcnn_r50_fpn_1x.py'\n",
+ "config_file = '../configs/faster_rcnn_r50_fpn_1x_coco.py'\n",
"# download the checkpoint from model zoo and put it in `checkpoints/`\n",
"checkpoint_file = '../checkpoints/faster_rcnn_r50_fpn_1x_20181010-3d1b3351.pth'"
]
@@ -85,8 +85,17 @@
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.3"
+ },
+ "pycharm": {
+ "stem_cell": {
+ "cell_type": "raw",
+ "source": [],
+ "metadata": {
+ "collapsed": false
+ }
+ }
}
},
"nbformat": 4,
"nbformat_minor": 2
-}
+}
\ No newline at end of file
diff --git a/docs/GETTING_STARTED.md b/docs/GETTING_STARTED.md
index 38f677c6597c1d009cb38a1c085f9f332338409a..55cd16595d4473448a6250ddd77d92ffe95dc12c 100644
--- a/docs/GETTING_STARTED.md
+++ b/docs/GETTING_STARTED.md
@@ -38,7 +38,7 @@ Assume that you have already downloaded the checkpoints to the directory `checkp
1. Test Faster R-CNN and visualize the results. Press any key for the next image.
```shell
-python tools/test.py configs/faster_rcnn_r50_fpn_1x.py \
+python tools/test.py configs/faster_rcnn_r50_fpn_1x_coco.py \
checkpoints/faster_rcnn_r50_fpn_1x_20181010-3d1b3351.pth \
--show
```
@@ -54,7 +54,7 @@ python tools/test.py configs/pascal_voc/faster_rcnn_r50_fpn_1x_voc.py \
3. Test Mask R-CNN with 8 GPUs, and evaluate the bbox and mask AP.
```shell
-./tools/dist_test.sh configs/mask_rcnn_r50_fpn_1x.py \
+./tools/dist_test.sh configs/mask_rcnn_r50_fpn_1x_coco.py \
checkpoints/mask_rcnn_r50_fpn_1x_20181010-069fa190.pth \
8 --out results.pkl --eval bbox segm
```
@@ -62,7 +62,7 @@ python tools/test.py configs/pascal_voc/faster_rcnn_r50_fpn_1x_voc.py \
4. Test Mask R-CNN on COCO test-dev with 8 GPUs, and generate the json file to be submit to the official evaluation server.
```shell
-./tools/dist_test.sh configs/mask_rcnn_r50_fpn_1x.py \
+./tools/dist_test.sh configs/mask_rcnn_r50_fpn_1x_coco.py \
checkpoints/mask_rcnn_r50_fpn_1x_20181010-069fa190.pth \
8 --format_only --options "jsonfile_prefix=./mask_rcnn_test-dev_results"
```
@@ -90,7 +90,7 @@ python demo/webcam_demo.py ${CONFIG_FILE} ${CHECKPOINT_FILE} [--device ${GPU_ID}
Examples:
```shell
-python demo/webcam_demo.py configs/faster_rcnn_r50_fpn_1x.py \
+python demo/webcam_demo.py configs/faster_rcnn_r50_fpn_1x_coco.py \
checkpoints/faster_rcnn_r50_fpn_1x_20181010-3d1b3351.pth
```
@@ -103,7 +103,7 @@ Here is an example of building the model and test given images.
from mmdet.apis import init_detector, inference_detector, show_result
import mmcv
-config_file = 'configs/faster_rcnn_r50_fpn_1x.py'
+config_file = 'configs/faster_rcnn_r50_fpn_1x_coco.py'
checkpoint_file = 'checkpoints/faster_rcnn_r50_fpn_1x_20181010-3d1b3351.pth'
# build the model from a config file and a checkpoint file
@@ -139,7 +139,7 @@ from mmdet.apis import init_detector, async_inference_detector, show_result
from mmdet.utils.contextmanagers import concurrent
async def main():
- config_file = 'configs/faster_rcnn_r50_fpn_1x.py'
+ config_file = 'configs/faster_rcnn_r50_fpn_1x_coco.py'
checkpoint_file = 'checkpoints/faster_rcnn_r50_fpn_1x_20181010-3d1b3351.pth'
device = 'cuda:0'
model = init_detector(config_file, checkpoint=checkpoint_file, device=device)
@@ -201,7 +201,7 @@ If you want to specify the working directory in the command, you can add an argu
Optional arguments are:
-- `--validate` (**strongly recommended**): Perform evaluation at every k (default value is 1, which can be modified like [this](https://github.com/open-mmlab/mmdetection/blob/master/configs/mask_rcnn_r50_fpn_1x.py#L174)) epochs during the training.
+- `--validate` (**strongly recommended**): Perform evaluation at every k (default value is 1, which can be modified like [this](https://github.com/open-mmlab/mmdetection/blob/master/configs/mask_rcnn_r50_fpn_1x_coco.py#L174)) epochs during the training.
- `--work_dir ${WORK_DIR}`: Override the working directory specified in the config file.
- `--resume_from ${CHECKPOINT_FILE}`: Resume from a previous checkpoint file.
@@ -220,7 +220,7 @@ If you run MMDetection on a cluster managed with [slurm](https://slurm.schedmd.c
Here is an example of using 16 GPUs to train Mask R-CNN on the dev partition.
```shell
-./tools/slurm_train.sh dev mask_r50_1x configs/mask_rcnn_r50_fpn_1x.py /nfs/xxxx/mask_rcnn_r50_fpn_1x 16
+./tools/slurm_train.sh dev mask_r50_1x configs/mask_rcnn_r50_fpn_1x_coco.py /nfs/xxxx/mask_rcnn_r50_fpn_1x 16
```
You can check [slurm_train.sh](https://github.com/open-mmlab/mmdetection/blob/master/tools/slurm_train.sh) for full arguments and environment variables.
diff --git a/mmdet/models/bbox_heads/__init__.py b/mmdet/models/bbox_heads/__init__.py
index a668bdb0182361b5213adb877a21246e23ab4ede..708f8ae643831d4e06a41f68aa63aec84b018b44 100644
--- a/mmdet/models/bbox_heads/__init__.py
+++ b/mmdet/models/bbox_heads/__init__.py
@@ -1,7 +1,9 @@
from .bbox_head import BBoxHead
-from .convfc_bbox_head import ConvFCBBoxHead, SharedFCBBoxHead
+from .convfc_bbox_head import (ConvFCBBoxHead, Shared2FCBBoxHead,
+ Shared4Conv1FCBBoxHead)
from .double_bbox_head import DoubleConvFCBBoxHead
__all__ = [
- 'BBoxHead', 'ConvFCBBoxHead', 'SharedFCBBoxHead', 'DoubleConvFCBBoxHead'
+ 'BBoxHead', 'ConvFCBBoxHead', 'Shared2FCBBoxHead',
+ 'Shared4Conv1FCBBoxHead', 'DoubleConvFCBBoxHead'
]
diff --git a/mmdet/models/bbox_heads/convfc_bbox_head.py b/mmdet/models/bbox_heads/convfc_bbox_head.py
index c43707da09e49ad6903712d0cb9a912055f8d96f..a30765e25b6eca922c40ca7f0c88662b97946d89 100644
--- a/mmdet/models/bbox_heads/convfc_bbox_head.py
+++ b/mmdet/models/bbox_heads/convfc_bbox_head.py
@@ -172,13 +172,28 @@ class ConvFCBBoxHead(BBoxHead):
@HEADS.register_module
-class SharedFCBBoxHead(ConvFCBBoxHead):
+class Shared2FCBBoxHead(ConvFCBBoxHead):
- def __init__(self, num_fcs=2, fc_out_channels=1024, *args, **kwargs):
- assert num_fcs >= 1
- super(SharedFCBBoxHead, self).__init__(
+ def __init__(self, fc_out_channels=1024, *args, **kwargs):
+ super(Shared2FCBBoxHead, self).__init__(
num_shared_convs=0,
- num_shared_fcs=num_fcs,
+ num_shared_fcs=2,
+ num_cls_convs=0,
+ num_cls_fcs=0,
+ num_reg_convs=0,
+ num_reg_fcs=0,
+ fc_out_channels=fc_out_channels,
+ *args,
+ **kwargs)
+
+
+@HEADS.register_module
+class Shared4Conv1FCBBoxHead(ConvFCBBoxHead):
+
+ def __init__(self, fc_out_channels=1024, *args, **kwargs):
+ super(Shared4Conv1FCBBoxHead, self).__init__(
+ num_shared_convs=4,
+ num_shared_fcs=1,
num_cls_convs=0,
num_cls_fcs=0,
num_reg_convs=0,
diff --git a/tests/async_benchmark.py b/tests/async_benchmark.py
index b4d885e203f89c67cfb8308f68b19feef177a535..d63c56bf18f17a6f7fc4fc310ec961385b7efbde 100644
--- a/tests/async_benchmark.py
+++ b/tests/async_benchmark.py
@@ -31,7 +31,8 @@ async def main():
"""
project_dir = os.path.abspath(os.path.dirname(os.path.dirname(__file__)))
- config_file = os.path.join(project_dir, 'configs/mask_rcnn_r50_fpn_1x.py')
+ config_file = os.path.join(project_dir,
+ 'configs/mask_rcnn_r50_fpn_1x_coco.py')
checkpoint_file = os.path.join(
project_dir, 'checkpoints/mask_rcnn_r50_fpn_1x_20181010-069fa190.pth')
diff --git a/tests/test_async.py b/tests/test_async.py
index 2733c7f5ef747dd03144aaac5d018256403dc60d..f14d03c5db0f6680baa320306f3b2e8ba237dd82 100644
--- a/tests/test_async.py
+++ b/tests/test_async.py
@@ -69,8 +69,8 @@ class AsyncInferenceTestCase(AsyncTestCase):
pytest.skip('test requires GPU and torch+cuda')
root_dir = os.path.dirname(os.path.dirname(__name__))
- model_config = os.path.join(root_dir,
- 'configs/mask_rcnn_r50_fpn_1x.py')
+ model_config = os.path.join(
+ root_dir, 'configs/mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py')
detector = MaskRCNNDetector(model_config)
await detector.init()
img_path = os.path.join(root_dir, 'demo/demo.jpg')
diff --git a/tests/test_config.py b/tests/test_config.py
index 599a1cf59a5dcddd7fe8250c743fd24604f9ce25..968b0ed3b03f91b59b1a1d60ff63bd0225c53df6 100644
--- a/tests/test_config.py
+++ b/tests/test_config.py
@@ -20,7 +20,7 @@ def test_config_build_detector():
"""
Test that all detection models defined in the configs can be initialized.
"""
- from xdoctest.utils import import_module_from_path
+ from mmcv import Config
from mmdet.models import build_detector
config_dpath = _get_config_directory()
@@ -33,128 +33,130 @@ def test_config_build_detector():
# Only tests a representative subset of configurations
config_names = [
- # 'dcn/faster_rcnn_dconv_c3-c5_r50_fpn_1x.py',
- # 'dcn/cascade_mask_rcnn_dconv_c3-c5_r50_fpn_1x.py',
- # 'dcn/faster_rcnn_dpool_r50_fpn_1x.py',
- 'dcn/mask_rcnn_dconv_c3-c5_r50_fpn_1x.py',
- # 'dcn/faster_rcnn_dconv_c3-c5_x101_32x4d_fpn_1x.py',
- # 'dcn/cascade_rcnn_dconv_c3-c5_r50_fpn_1x.py',
- # 'dcn/faster_rcnn_mdpool_r50_fpn_1x.py',
- # 'dcn/faster_rcnn_mdconv_c3-c5_group4_r50_fpn_1x.py',
- # 'dcn/faster_rcnn_mdconv_c3-c5_r50_fpn_1x.py',
+ # 'dcn/faster_rcnn_dconv_c3-c5_r50_fpn_1x_coco.py',
+ # 'dcn/cascade_mask_rcnn_dconv_c3-c5_r50_fpn_1x_coco.py',
+ # 'dcn/faster_rcnn_dpool_r50_fpn_1x_coco.py',
+ 'dcn/mask_rcnn_r50_fpn_dconv_c3-c5_1x_coco.py',
+ # 'dcn/faster_rcnn_dconv_c3-c5_x101_32x4d_fpn_1x_coco.py',
+ # 'dcn/cascade_rcnn_dconv_c3-c5_r50_fpn_1x_coco.py',
+ # 'dcn/faster_rcnn_mdpool_r50_fpn_1x_coco.py',
+ # 'dcn/faster_rcnn_mdconv_c3-c5_group4_r50_fpn_1x_coco.py',
+ # 'dcn/faster_rcnn_mdconv_c3-c5_r50_fpn_1x_coco.py',
# ---
- # 'htc/htc_x101_32x4d_fpn_20e_16gpu.py',
- 'htc/htc_without_semantic_r50_fpn_1x.py',
- # 'htc/htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e.py',
- # 'htc/htc_x101_64x4d_fpn_20e_16gpu.py',
- # 'htc/htc_r50_fpn_1x.py',
+ # 'htc/htc_x101_32x4d_fpn_20e_16gpu_coco.py',
+ 'htc/htc_without_semantic_r50_fpn_1x_coco.py',
+ # 'htc/htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_coco.py',
+ # 'htc/htc_x101_64x4d_fpn_20e_16gpu_coco.py',
+ # 'htc/htc_r50_fpn_1x_coco.py',
# 'htc/htc_r101_fpn_20e.py',
# 'htc/htc_r50_fpn_20e.py',
# ---
'cityscapes/mask_rcnn_r50_fpn_1x_cityscapes.py',
# 'cityscapes/faster_rcnn_r50_fpn_1x_cityscapes.py',
# ---
- # 'scratch/scratch_faster_rcnn_r50_fpn_gn_6x.py',
- # 'scratch/scratch_mask_rcnn_r50_fpn_gn_6x.py',
+ # 'scratch/scratch_faster_rcnn_r50_fpn_gn_6x_coco.py',
+ # 'scratch/scratch_mask_rcnn_r50_fpn_gn_6x_coco.py',
# ---
- # 'grid_rcnn/grid_rcnn_gn_head_x101_32x4d_fpn_2x.py',
- 'grid_rcnn/grid_rcnn_gn_head_r50_fpn_2x.py',
+ # 'grid_rcnn/grid_rcnn_gn_head_x101_32x4d_fpn_2x_coco.py',
+ 'grid_rcnn/grid_rcnn_r50_fpn_gn-head_2x_coco.py',
# ---
- 'double_heads/dh_faster_rcnn_r50_fpn_1x.py',
+ 'double_heads/dh_faster_rcnn_r50_fpn_1x_coco.py',
# ---
- 'empirical_attention/faster_rcnn_r50_fpn_attention_0010_dcn_1x.py',
- # 'empirical_attention/faster_rcnn_r50_fpn_attention_1111_1x.py',
- # 'empirical_attention/faster_rcnn_r50_fpn_attention_0010_1x.py',
- # 'empirical_attention/faster_rcnn_r50_fpn_attention_1111_dcn_1x.py',
+ 'empirical_attention/faster_rcnn_r50_fpn_attention_0010_dcn_1x_coco'
+ '.py',
+ # 'empirical_attention/faster_rcnn_r50_fpn_attention_1111_1x_coco.py',
+ # 'empirical_attention/faster_rcnn_r50_fpn_attention_0010_1x_coco.py',
+ # 'empirical_attention/faster_rcnn_r50_fpn_attention_1111_dcn_1x_coco
+ # .py',
# ---
- # 'ms_rcnn/ms_rcnn_r101_caffe_fpn_1x.py',
- # 'ms_rcnn/ms_rcnn_x101_64x4d_fpn_1x.py',
- # 'ms_rcnn/ms_rcnn_r50_caffe_fpn_1x.py',
+ # 'ms_rcnn/ms_rcnn_r101_caffe_fpn_1x_coco.py',
+ # 'ms_rcnn/ms_rcnn_x101_64x4d_fpn_1x_coco.py',
+ # 'ms_rcnn/ms_rcnn_r50_caffe_fpn_1x_coco.py',
# ---
- # 'guided_anchoring/ga_faster_x101_32x4d_fpn_1x.py',
- # 'guided_anchoring/ga_rpn_x101_32x4d_fpn_1x.py',
- # 'guided_anchoring/ga_retinanet_r50_caffe_fpn_1x.py',
- # 'guided_anchoring/ga_fast_r50_caffe_fpn_1x.py',
- # 'guided_anchoring/ga_retinanet_x101_32x4d_fpn_1x.py',
- # 'guided_anchoring/ga_rpn_r101_caffe_rpn_1x.py',
- # 'guided_anchoring/ga_faster_r50_caffe_fpn_1x.py',
- 'guided_anchoring/ga_rpn_r50_caffe_fpn_1x.py',
+ # 'guided_anchoring/ga_faster_x101_32x4d_fpn_1x_coco.py',
+ # 'guided_anchoring/ga_rpn_x101_32x4d_fpn_1x_coco.py',
+ # 'guided_anchoring/ga_retinanet_r50_caffe_fpn_1x_coco.py',
+ # 'guided_anchoring/ga_fast_r50_caffe_fpn_1x_coco.py',
+ # 'guided_anchoring/ga_retinanet_x101_32x4d_fpn_1x_coco.py',
+ # 'guided_anchoring/ga_rpn_r101_caffe_rpn_1x_coco.py',
+ # 'guided_anchoring/ga_faster_r50_caffe_fpn_1x_coco.py',
+ 'guided_anchoring/ga_rpn_r50_caffe_fpn_1x_coco.py',
# ---
- 'foveabox/fovea_r50_fpn_4gpu_1x.py',
- # 'foveabox/fovea_align_gn_ms_r101_fpn_4gpu_2x.py',
- # 'foveabox/fovea_align_gn_r50_fpn_4gpu_2x.py',
- # 'foveabox/fovea_align_gn_r101_fpn_4gpu_2x.py',
- 'foveabox/fovea_align_gn_ms_r50_fpn_4gpu_2x.py',
+ 'foveabox/fovea_r50_fpn_4x4_1x_coco.py',
+ # 'foveabox/fovea_align_gn_ms_r101_fpn_4gpu_2x_coco.py',
+ # 'foveabox/fovea_align_gn_r50_fpn_4gpu_2x_coco.py',
+ # 'foveabox/fovea_align_gn_r101_fpn_4gpu_2x_coco.py',
+ 'foveabox/fovea_align_r50_fpn_gn-head_mstrain_640-800_4x4_2x_coco.py',
# ---
- # 'hrnet/cascade_rcnn_hrnetv2p_w32_20e.py',
- # 'hrnet/mask_rcnn_hrnetv2p_w32_1x.py',
- # 'hrnet/cascade_mask_rcnn_hrnetv2p_w32_20e.py',
- # 'hrnet/htc_hrnetv2p_w32_20e.py',
- # 'hrnet/faster_rcnn_hrnetv2p_w18_1x.py',
- # 'hrnet/mask_rcnn_hrnetv2p_w18_1x.py',
- # 'hrnet/faster_rcnn_hrnetv2p_w32_1x.py',
- # 'hrnet/faster_rcnn_hrnetv2p_w40_1x.py',
- 'hrnet/fcos_hrnetv2p_w32_gn_1x_4gpu.py',
+ # 'hrnet/cascade_rcnn_hrnetv2p_w32_20e_coco.py',
+ # 'hrnet/mask_rcnn_hrnetv2p_w32_1x_coco.py',
+ # 'hrnet/cascade_mask_rcnn_hrnetv2p_w32_20e_coco.py',
+ # 'hrnet/htc_hrnetv2p_w32_20e_coco.py',
+ # 'hrnet/faster_rcnn_hrnetv2p_w18_1x_coco.py',
+ # 'hrnet/mask_rcnn_hrnetv2p_w18_1x_coco.py',
+ # 'hrnet/faster_rcnn_hrnetv2p_w32_1x_coco.py',
+ # 'hrnet/faster_rcnn_hrnetv2p_w40_1x_coco.py',
+ 'hrnet/fcos_hrnetv2p_w32_gn-head_4x4_1x_coco.py',
# ---
- # 'gn+ws/faster_rcnn_r50_fpn_gn_ws_1x.py',
- # 'gn+ws/mask_rcnn_x101_32x4d_fpn_gn_ws_2x.py',
- 'gn+ws/mask_rcnn_r50_fpn_gn_ws_2x.py',
- # 'gn+ws/mask_rcnn_r50_fpn_gn_ws_20_23_24e.py',
+ # 'gn+ws/faster_rcnn_r50_fpn_gn_ws_1x_coco.py',
+ # 'gn+ws/mask_rcnn_x101_32x4d_fpn_gn_ws_2x_coco.py',
+ 'gn+ws/mask_rcnn_r50_fpn_gn_ws-all_2x_coco.py',
+ # 'gn+ws/mask_rcnn_r50_fpn_gn_ws_20_23_24e_coco.py',
# ---
# 'wider_face/ssd300_wider_face.py',
# ---
- 'pascal_voc/ssd300_voc.py',
+ 'pascal_voc/ssd300_voc0712.py',
'pascal_voc/faster_rcnn_r50_fpn_1x_voc0712.py',
- 'pascal_voc/ssd512_voc.py',
+ 'pascal_voc/ssd512_voc0712.py',
# ---
- # 'gcnet/mask_rcnn_r4_gcb_c3-c5_r50_fpn_syncbn_1x.py',
- # 'gcnet/mask_rcnn_r16_gcb_c3-c5_r50_fpn_syncbn_1x.py',
- # 'gcnet/mask_rcnn_r4_gcb_c3-c5_r50_fpn_1x.py',
- # 'gcnet/mask_rcnn_r16_gcb_c3-c5_r50_fpn_1x.py',
- 'gcnet/mask_rcnn_r50_fpn_sbn_1x.py',
+ # 'gcnet/mask_rcnn_r4_gcb_c3-c5_r50_fpn_syncbn_1x_coco.py',
+ # 'gcnet/mask_rcnn_r16_gcb_c3-c5_r50_fpn_syncbn_1x_coco.py',
+ # 'gcnet/mask_rcnn_r4_gcb_c3-c5_r50_fpn_1x_coco.py',
+ # 'gcnet/mask_rcnn_r16_gcb_c3-c5_r50_fpn_1x_coco.py',
+ 'gcnet/mask_rcnn_r50_fpn_syncbn-backbone_1x_coco.py',
# ---
- 'gn/mask_rcnn_r50_fpn_gn_contrib_2x.py',
- # 'gn/mask_rcnn_r50_fpn_gn_2x.py',
- # 'gn/mask_rcnn_r101_fpn_gn_2x.py',
+ 'gn/mask_rcnn_r50_fpn_gn-all_contrib_2x_coco.py',
+ # 'gn/mask_rcnn_r50_fpn_gn_2x_coco.py',
+ # 'gn/mask_rcnn_r101_fpn_gn_2x_coco.py',
# ---
# 'reppoints/reppoints_moment_x101_dcn_fpn_2x.py',
- 'reppoints/reppoints_moment_r50_fpn_2x.py',
- # 'reppoints/reppoints_moment_x101_dcn_fpn_2x_mt.py',
- 'reppoints/reppoints_partial_minmax_r50_fpn_1x.py',
- 'reppoints/bbox_r50_grid_center_fpn_1x.py',
- # 'reppoints/reppoints_moment_r101_dcn_fpn_2x.py',
- # 'reppoints/reppoints_moment_r101_fpn_2x_mt.py',
- # 'reppoints/reppoints_moment_r50_fpn_2x_mt.py',
- 'reppoints/reppoints_minmax_r50_fpn_1x.py',
- # 'reppoints/reppoints_moment_r50_fpn_1x.py',
- # 'reppoints/reppoints_moment_r101_fpn_2x.py',
- # 'reppoints/reppoints_moment_r101_dcn_fpn_2x_mt.py',
- 'reppoints/bbox_r50_grid_fpn_1x.py',
+ 'reppoints/reppoints_moment_r50_fpn_gn-neck+head_2x_coco.py',
+ # 'reppoints/reppoints_moment_x101_dcn_fpn_2x_mt_coco.py',
+ 'reppoints/reppoints_partial_minmax_r50_fpn_gn-neck+head_1x_coco.py',
+ 'reppoints/bbox_r50_grid_center_fpn_gn-neck+head_1x_coco.py',
+ # 'reppoints/reppoints_moment_r101_dcn_fpn_2x_coco.py',
+ # 'reppoints/reppoints_moment_r101_fpn_2x_mt_coco.py',
+ # 'reppoints/reppoints_moment_r50_fpn_2x_mt_coco.py',
+ 'reppoints/reppoints_minmax_r50_fpn_gn-neck+head_1x_coco.py',
+ # 'reppoints/reppoints_moment_r50_fpn_1x_coco.py',
+ # 'reppoints/reppoints_moment_r101_fpn_2x_coco.py',
+ # 'reppoints/reppoints_moment_r101_dcn_fpn_2x_mt_coco.py',
+ 'reppoints/bbox_r50_grid_fpn_gn-neck+head_1x_coco.py',
# ---
- # 'fcos/fcos_mstrain_640_800_x101_64x4d_fpn_gn_2x.py',
- # 'fcos/fcos_mstrain_640_800_r101_caffe_fpn_gn_2x_4gpu.py',
- 'fcos/fcos_r50_caffe_fpn_gn_1x_4gpu.py',
+ # 'fcos/fcos_mstrain_640_800_x101_64x4d_fpn_gn_2x_coco.py',
+ # 'fcos/fcos_mstrain_640_800_r101_caffe_fpn_gn_2x_4gpu_coco.py',
+ 'fcos/fcos_r50_caffe_fpn_gn-head_4x4_1x_coco.py',
# ---
- 'albu_example/mask_rcnn_r50_fpn_1x.py',
+ 'albu_example/mask_rcnn_r50_fpn_albu_1x_coco.py',
# ---
- 'libra_rcnn/libra_faster_rcnn_r50_fpn_1x.py',
- # 'libra_rcnn/libra_retinanet_r50_fpn_1x.py',
- # 'libra_rcnn/libra_faster_rcnn_r101_fpn_1x.py',
- # 'libra_rcnn/libra_faster_rcnn_x101_64x4d_fpn_1x.py',
- # 'libra_rcnn/libra_fast_rcnn_r50_fpn_1x.py',
+ 'libra_rcnn/libra_faster_rcnn_r50_fpn_1x_coco.py',
+ # 'libra_rcnn/libra_retinanet_r50_fpn_1x_coco.py',
+ # 'libra_rcnn/libra_faster_rcnn_r101_fpn_1x_coco.py',
+ # 'libra_rcnn/libra_faster_rcnn_x101_64x4d_fpn_1x_coco.py',
+ # 'libra_rcnn/libra_fast_rcnn_r50_fpn_1x_coco.py',
# ---
- # 'ghm/retinanet_ghm_r50_fpn_1x.py',
+ # 'ghm/retinanet_ghm_r50_fpn_1x_coco.py',
# ---
- # 'fp16/retinanet_r50_fpn_fp16_1x.py',
- 'fp16/mask_rcnn_r50_fpn_fp16_1x.py',
- 'fp16/faster_rcnn_r50_fpn_fp16_1x.py'
+ # 'fp16/retinanet_r50_fpn_fp16_1x_coco.py',
+ 'fp16/mask_rcnn_r50_fpn_fp16_1x_coco.py',
+ 'fp16/faster_rcnn_r50_fpn_fp16_1x_coco.py'
]
print('Using {} config files'.format(len(config_names)))
for config_fname in config_names:
config_fpath = join(config_dpath, config_fname)
- config_mod = import_module_from_path(config_fpath)
+ config_mod = Config.fromfile(config_fpath)
config_mod.model
config_mod.train_cfg
@@ -178,7 +180,7 @@ def test_config_data_pipeline():
CommandLine:
xdoctest -m tests/test_config.py test_config_build_data_pipeline
"""
- from xdoctest.utils import import_module_from_path
+ from mmcv import Config
from mmdet.datasets.pipelines import Compose
import numpy as np
@@ -189,17 +191,17 @@ def test_config_data_pipeline():
# TODO: test pipelines using Albu, current Albu throw None given empty GT
config_names = [
'wider_face/ssd300_wider_face.py',
- 'pascal_voc/ssd300_voc.py',
- 'pascal_voc/ssd512_voc.py',
+ 'pascal_voc/ssd300_voc0712.py',
+ 'pascal_voc/ssd512_voc0712.py',
# 'albu_example/mask_rcnn_r50_fpn_1x.py',
- 'fp16/mask_rcnn_r50_fpn_fp16_1x.py',
+ 'fp16/mask_rcnn_r50_fpn_fp16_1x_coco.py',
]
print('Using {} config files'.format(len(config_names)))
for config_fname in config_names:
config_fpath = join(config_dpath, config_fname)
- config_mod = import_module_from_path(config_fpath)
+ config_mod = Config.fromfile(config_fpath)
# remove loading pipeline
loading_pipeline = config_mod.train_pipeline.pop(0)
diff --git a/tests/test_forward.py b/tests/test_forward.py
index 5ba56bf24ff51861e3e897a34aa526d702d520a2..2599fb115b92b44bf98e28ff89b6576e5df80bb1 100644
--- a/tests/test_forward.py
+++ b/tests/test_forward.py
@@ -27,10 +27,10 @@ def _get_config_module(fname):
"""
Load a configuration as a python module
"""
- from xdoctest.utils import import_module_from_path
+ from mmcv import Config
config_dpath = _get_config_directory()
config_fpath = join(config_dpath, fname)
- config_mod = import_module_from_path(config_fpath)
+ config_mod = Config.fromfile(config_fpath)
return config_mod
@@ -48,7 +48,7 @@ def _get_detector_cfg(fname):
def test_ssd300_forward():
- model, train_cfg, test_cfg = _get_detector_cfg('ssd300_coco.py')
+ model, train_cfg, test_cfg = _get_detector_cfg('ssd/ssd300_coco.py')
model['pretrained'] = None
from mmdet.models import build_detector
@@ -82,7 +82,8 @@ def test_ssd300_forward():
def test_rpn_forward():
- model, train_cfg, test_cfg = _get_detector_cfg('rpn_r50_fpn_1x.py')
+ model, train_cfg, test_cfg = _get_detector_cfg(
+ 'rpn/rpn_r50_fpn_1x_coco.py')
model['pretrained'] = None
from mmdet.models import build_detector
@@ -112,7 +113,7 @@ def test_rpn_forward():
def test_retina_ghm_forward():
model, train_cfg, test_cfg = _get_detector_cfg(
- 'ghm/retinanet_ghm_r50_fpn_1x.py')
+ 'ghm/retinanet_ghm_r50_fpn_1x_coco.py')
model['pretrained'] = None
from mmdet.models import build_detector
@@ -176,7 +177,7 @@ def test_cascade_forward():
raise pytest.skip('requires torchvision on cpu')
model, train_cfg, test_cfg = _get_detector_cfg(
- 'cascade_rcnn_r50_fpn_1x.py')
+ 'cascade_rcnn/cascade_rcnn_r50_fpn_1x_coco.py')
model['pretrained'] = None
# torchvision roi align supports CPU
model['bbox_roi_extractor']['roi_layer']['use_torchvision'] = True
@@ -228,7 +229,8 @@ def test_faster_rcnn_forward():
import pytest
raise pytest.skip('requires torchvision on cpu')
- model, train_cfg, test_cfg = _get_detector_cfg('faster_rcnn_r50_fpn_1x.py')
+ model, train_cfg, test_cfg = _get_detector_cfg(
+ 'faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py')
model['pretrained'] = None
# torchvision roi align supports CPU
model['bbox_roi_extractor']['roi_layer']['use_torchvision'] = True
@@ -281,7 +283,7 @@ def test_faster_rcnn_ohem_forward():
raise pytest.skip('requires torchvision on cpu')
model, train_cfg, test_cfg = _get_detector_cfg(
- 'faster_rcnn_ohem_r50_fpn_1x.py')
+ 'faster_rcnn/faster_rcnn_r50_fpn_ohem_1x_coco.py')
model['pretrained'] = None
# torchvision roi align supports CPU
model['bbox_roi_extractor']['roi_layer']['use_torchvision'] = True
diff --git a/tests/test_heads.py b/tests/test_heads.py
index b1e4ceebf45d284e5a96731fe247abaebd6fe242..e835c9e391e4e3697179cd43088e0be216b2d540 100644
--- a/tests/test_heads.py
+++ b/tests/test_heads.py
@@ -241,7 +241,7 @@ def test_refine_boxes():
'rng': 699807
},
- # Corner case: similar num rois and images
+ # Corner case: cal_similarity num rois and images
{
'n_roi': 20,
'n_img': 20,
diff --git a/tests/test_sampler.py b/tests/test_sampler.py
index c75360268e6fd3afd6cbfb4ecf09a4de5cbf138a..cdae8f4821cfc929e779f55077bc929a7e113818 100644
--- a/tests/test_sampler.py
+++ b/tests/test_sampler.py
@@ -100,7 +100,7 @@ def _context_for_ohem():
sys.path.insert(0, dirname(__file__))
from test_forward import _get_detector_cfg
model, train_cfg, test_cfg = _get_detector_cfg(
- 'faster_rcnn_ohem_r50_fpn_1x.py')
+ 'faster_rcnn/faster_rcnn_r50_fpn_ohem_1x_coco.py')
model['pretrained'] = None
# torchvision roi align supports CPU
model['bbox_roi_extractor']['roi_layer']['use_torchvision'] = True
diff --git a/tools/train.py b/tools/train.py
index 01df1caef34098041fc080d1a5e4a336ff45a62a..9ce7fa2f4a64772b54a97a2eb269613c67d13f83 100644
--- a/tools/train.py
+++ b/tools/train.py
@@ -65,6 +65,10 @@ def main():
# update configs according to CLI args
if args.work_dir is not None:
cfg.work_dir = args.work_dir
+ elif cfg.get('work_dir', None) is None:
+ # use filename as default work_dir
+ cfg.work_dir = osp.join('./work_dirs',
+ osp.splitext(osp.basename(args.config))[0])
if args.resume_from is not None:
cfg.resume_from = args.resume_from
cfg.gpus = args.gpus