add mimic fpn B->T

configs/distillers/mimic_fpn/mb_cascade_rcnn_swinB_fpn_3x_distill_cascade_rcnn_swinT_fpn_1x_coco.py

+_base_ = [
+    '../../_base_/models/mask_rcnn_swin_fpn.py',
+    '../../_base_/datasets/coco_instance.py',
+    '../../_base_/schedules/schedule_1x.py', '../../_base_/default_runtime.py'
+]
+
+# model settings
+find_unused_parameters=True
+weight=1
+distiller = dict(
+    type='FPNDistiller',
+    teacher_pretrained = '/mnt/data3/wangguohua/model/mmdet/swin/cascade_mask_rcnn_swin_base_patch4_window7.pth',
+    init_student = True,
+    distill_cfg = [ dict(feature_level = 0,
+                         methods=[dict(type='MSELoss',
+                                       name='loss_mb_0',
+                                       student_channels = 256,
+                                       teacher_channels = 256,
+                                       weight = weight,
+                                       )
+                                ]
+                        ),
+                    dict(feature_level = 1,
+                         methods=[dict(type='MSELoss',
+                                       name='loss_mb_1',
+                                       student_channels = 256,
+                                       teacher_channels = 256,
+                                       weight = weight,
+                                       )
+                                ]
+                        ),
+                    dict(feature_level = 2,
+                         methods=[dict(type='MSELoss',
+                                       name='loss_mb_2',
+                                       student_channels = 256,
+                                       teacher_channels = 256,
+                                       weight = weight,
+                                       )
+                                ]
+                        ),
+                    dict(feature_level = 3,
+                         methods=[dict(type='MSELoss',
+                                       name='loss_mb_3',
+                                       student_channels = 256,
+                                       teacher_channels = 256,
+                                       weight = weight,
+                                       )
+                                ]
+                        ),
+                   ]
+    )
+
+student_cfg = 'configs/swin/cascade_mask_rcnn_swin_tiny_patch4_window7_mstrain_480-800_giou_4conv1f_adamw_1x_coco.py'
+teacher_cfg = 'configs/swin/cascade_mask_rcnn_swin_base_patch4_window7_mstrain_480-800_giou_4conv1f_adamw_3x_coco.py'
+
+data = dict(
+    samples_per_gpu=2,
+    workers_per_gpu=2,)
+#data = dict(train=dict(pipeline=train_pipeline))
+
+optimizer = dict(_delete_=True, type='AdamW', lr=0.0001, betas=(0.9, 0.999), weight_decay=0.05,
+                 paramwise_cfg=dict(custom_keys={'absolute_pos_embed': dict(decay_mult=0.),
+                                                 'relative_position_bias_table': dict(decay_mult=0.),
+                                                 'norm': dict(decay_mult=0.)}))
+lr_config = dict(step=[8, 11])
+#runner = dict(type='EpochBasedRunnerAmp', max_epochs=12)
+runner = dict(type='EpochBasedRunner', max_epochs=12)
+# do not use mmdet version fp16
+# fp16 = None
+# optimizer_config = dict(
+#     type="DistOptimizerHook",
+#     update_interval=1,
+#     grad_clip=None,
+#     coalesce=True,
+#     bucket_size_mb=-1,
+#     use_fp16=True,
+# )
\ No newline at end of file