Unable to find a valid cuDNN algorithm

File “/home/hong/.conda/envs/Ner2Ner/lib/python3.7/site-packages/torch/nn/modules/conv.py”, line 790, in forward
output_padding, self.groups, self.dilation)
RuntimeError: Unable to find a valid cuDNN algorithm to run convolution

i’m running this repo : GitHub - TaoHuang2018/Neighbor2Neighbor: Neighbor2Neighbor: Self-Supervised Denoising from Single Noisy Images
on ubuntu22.04 pytorch1.5 with cuda: torch==1.5.0+cu101 torchvision==0.6.0+cu101
but i get the above error
could any one give some advice to fix this?

Could you update to the latest stable or nightly PyTorch release and check if you would still run into the error, please?