Update no_grad usage to inference_mode if possible #2534
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
rwightman
reviewed
Jul 3, 2025
train.py
Outdated
| @@ -512,7 +512,7 @@ def main(): | |||
| **args.model_kwargs, | |||
| ) | |||
| if args.head_init_scale is not None: | |||
| with torch.no_grad(): | |||
| with torch.inference_mode(): | |||
There was a problem hiding this comment.
This one should probably remain as no_grad
There was a problem hiding this comment.
I think it works. It does not interfer with the graph / properties of existing variable, and the mul just scale and does not create new params.
I just tried to run train.py for some steps with those params and saw no error
--model resnet18 --dataset hfds/imagenet-1k --head-init-scale .1
There was a problem hiding this comment.
All the same, would like to revert that to previous... everything else looks good assuming no test fails, which I doubt.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I updated no_grad references in the repository to inference_mode for some slight performance gain.
References for no_grad / inference_mode difference:
https://discuss.pytorch.org/t/pytorch-torch-no-grad-vs-torch-inference-mode/134099/2?u=timgianitsos
https://stackoverflow.com/questions/74191070/why-are-there-two-different-flags-to-disable-gradient-computation-in-pytorch/74197846#74197846