lishuang

https://github.com/open-mmlab/mmcv/blob/48d990258549ca626fcf8c34488c00ed6fce108a/tests/test_ops/test_onnx.py#L150

in mmcv,there is RoIAlign code for cuda,onnx and pytorch. As @irvingzhang0512 said,you can build specific plugins to support RoIAlign. And the link is the code about the RoIAlign. You can...

@vivva You can refer [this](https://github.com/open-mmlab/mmaction2/tree/master/demo#spatiotemporal-action-detection-video-demo) In demo folder ,there already was a SpatioTemporal Action Detection Video Demo

@ahkarami Where is the the detected action boundary in the temporal action detection?

@ahkarami I know what you mean. You're right.

``` -- -- ******** Summary ******** -- CMake version : 3.21.0 -- CMake command : /workspace/cmake-3.21.0-linux-x86_64/bin/cmake -- System : Linux -- C++ compiler : /usr/bin/c++ -- C++ compiler version :...

[onnxruntime](https://github.com/microsoft/onnxruntime/tree/rel-1.10.0)

目前已在三个模型上实现了剪枝操作。 稀疏度不够吧，觉得数据集不行可以试试coco和voc的效果

tanluren/yolov3-channel-and-layer-pruning里面有参数解释

训练时部分loss值为nan，建议debug调试