In old version of STGCN,I got a error ''RuntimeError: Given groups=1, weight of size [400, 256, 1, 1], expected input[64, 3, 1, 1] to have 256 channels, but got 3 channels instead''

[KrisLee512]

@yysijie Firstly,I thanks for your prominent work. I download the old version of STGCN,when I run the command ''python main.py recognition -c config/st_gcn/kinetics-skeleton/train.yaml'' according to the OLD_README.md,I got the error as the title said. My sincere thanks.

[a726]

@yysijie Firstly,I thanks for your prominent work. I download the old version of STGCN,when I run the command ''python main.py recognition -c config/st_gcn/kinetics-skeleton/train.yaml'' according to the OLD_README.md,I got the error as the title said. My sincere thanks.

hi，i meet same problem，can you solve it? can you help me.

[pranavgundewar]

Hi! Please roll back to PyTorch 1.2 or use a single GPU. That should do the trick.

[KrisLee512]

thanks, according  to your suggestion, it works well发自我的华为手机-------- 原始邮件 --------发件人： Pranav Gundewar [email protected]日期： 2020年6月18日周四 06:45收件人： open-mmlab/mmskeleton [email protected]抄送： KrisLee512 [email protected], Author [email protected]主 题： Re: [open-mmlab/mmskeleton] In old version of STGCN,I got a error ''RuntimeError: Given groups=1, weight of size [400, 256, 1, 1], expected input[64, 3, 1, 1] to have 256 channels, but got 3 channels instead'' (#323) Hi! Please roll back to PyTorch 1.2 or use a single GPU. That should do the trick.

—You are receiving this because you authored the thread.Reply to this email directly, view it on GitHub, or unsubscribe.

[huangjun12]

the same problem, with torch version 1.9.0