IGNN icon indicating copy to clipboard operation
IGNN copied to clipboard

How to run the codes on a computer with a nvidia 3090 and pytorch 1.7.1

Open weihancug opened this issue 3 years ago • 1 comments

Hello, there is a newly computer in my office, which is equipped with a nvidia GPU 3090. Therefore, Cuda 11 and Pytorch 1.7.1 are supported in this computer. We try to configure the environment and run the model, but we encounter such a problem,

"Legacy autograd function with non-static forward method is deprecated. "RuntimeError: Legacy autograd function with non-static forward method is deprecated. Please use new-style autograd function with static forward method. (Example: https://pytorch.org/docs/stable/autograd.html#torch.autograd.Function)

Do you know how to address the problem? Thanks for your time and kind assistance.

weihancug avatar Jan 06 '21 13:01 weihancug

Hello, there is a newly computer in my office, which is equipped with a nvidia GPU 3090. Therefore, Cuda 11 and Pytorch 1.7.1 are supported in this computer. We try to configure the environment and run the model, but we encounter such a problem,

"Legacy autograd function with non-static forward method is deprecated. "RuntimeError: Legacy autograd function with non-static forward method is deprecated. Please use new-style autograd function with static forward method. (Example: https://pytorch.org/docs/stable/autograd.html#torch.autograd.Function)

Do you know how to address the problem? Thanks for your time and kind assistance.

hello, I have the same problem. I have solved it by this: change the line 252 of "im2col.py" which is located in "/usr/local/lib/python3.6/dist-packages/pyinn/im2col.py" Col2Im(kernel_size, stride, padding)(input) toCol2Im(kernel_size, stride, padding).forward(input) and the line 244: Im2Col(kernel_size, stride, padding)(input) to Im2Col(kernel_size, stride, padding).forward(input)

hnsywangxin avatar Jun 29 '21 10:06 hnsywangxin