The problem when running the baseline FIM about nnr HOT 2 CLOSED

This is because Pytorch selects the slow 3D-CNN operator. To solve this problem, please modify the code in config.py (line 108 and 109):

from
torch.backends.cudnn.benchmark = False torch.backends.cudnn.deterministic = True
to
torch.backends.cudnn.benchmark = True torch.backends.cudnn.deterministic = False

I set the cudnn.benchmark and cudnn.deterministic to make all the experiment results reproducible. For FIM, you can modify the above options.

from nnr.

The problem is solved. Thanks for replying.

from nnr.