a question

[Youzebin]

when i try to train your great project, something wrong. do you know the reason for RuntimeError: No such operator aten::cudnn_convolution_backward_weight??? thank you very much~

[woctezuma]

• Error Running Demo #23

[Youzebin]

• Error Running Demo #23

I have another question, you mentioned in your paper that it is best to set the batch_size to 2048 when the resolution is between 16 and 64, but when I set the batch_size to 32, 2 gpu's and a separate batch_size of 16, it already shows that a single GPU with 16 video memory is not enough.
Is it possible that the batch_size you mentioned in your paper requires too much video memory?
@woctezuma

[xl-sr]

so you managed to solve your problem above?

[woctezuma]

Is it possible that the batch_size you mentioned in your paper requires too much video memory?

I am not an author of the paper.

By the way, if the issue is solved, feel free to create another issue for the new one.

[Youzebin]

so you managed to solve your problem above?

yes. I do some experiment to solve this problem.
when i use pytorch1.7.1, i solve the above problem.