Alvin

I found the cause of the problem: out[:, :-1, :fold] = x[:, 1:, :fold] # shift left out[:, 1:, fold: 2 * fold] = x[:, :-1, fold: 2 * fold]...

> Maybe you can try another implementation of shift module > > https://github.com/open-mmlab/mmaction2/blob/master/mmaction/models/backbones/resnet_tsm.py#L40-L121 @irvingzhang0512 Thanks ! It works now

同问 @Zhou-sx 请问gst下怎样实现mpp 硬解

导出onnx 正常 在1.4.0 下推理也正常

@Caesar-github 请问 什么原因呢

@lenka844 Hi~ Have you solved this problem?

I also used random mask render on the 360K dataset

@sangtv9 Hey! Has your problem been solved? whta is your backbone network? I have the same problem (backbone:mobileface)