From 6adc01ed7da2ffa96fd8b20c95047041c7196f37 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Haian=20Huang=28=E6=B7=B1=E5=BA=A6=E7=9C=B8=29?= <1286304229@qq.com> Date: Sat, 9 Jan 2021 12:52:46 +0800 Subject: [PATCH] Add cpu compatibility info in documentation (#4405) * Add cpu compatibility info * Add table about not support CPU only model info * Update docs about not support CPU only model info --- docs/get_started.md | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-) diff --git a/docs/get_started.md b/docs/get_started.md index faa534a3..19a1ad8d 100644 --- a/docs/get_started.md +++ b/docs/get_started.md @@ -118,12 +118,28 @@ In CPU mode you can run the demo/webcam_demo.py for example. However some functionality is gone in this mode: - Deformable Convolution +- Modulated Deformable Convolution +- ROI pooling - Deformable ROI pooling - CARAFE: Content-Aware ReAssembly of FEatures +- SyncBatchNorm +- CrissCrossAttention: Criss-Cross Attention +- MaskedConv2d +- Temporal Interlace Shift - nms_cuda - sigmoid_focal_loss_cuda +- bbox_overlaps -So if you try to run inference with a model containing deformable convolution you will get an error. +So if you try to run inference with a model containing above ops you will get an error. The following table lists the related methods that cannot inference on CPU due to dependency on these operators + +| Operator | Model | +| :-----------------------------------------------------: | :----------------------------------------------------------: | +| Deformable Convolution/Modulated Deformable Convolution | DCN銆丟uided Anchoring銆丷epPoints銆丆entripetalNet銆乂FNet銆丆ascadeRPN銆丯AS-FCOS銆丏etectoRS | +| MaskedConv2d | Guided Anchoring | +| CARAFE | CARAFE | +| SyncBatchNorm | ResNeSt | + +**Notice**: MMDetection does not support training with CPU for now. ### Another option: Docker Image -- GitLab