Restore the previous autograd setting#49
Open
martenlienen wants to merge 1 commit into
Open
Conversation
Indiscriminately enabling autograd at the end of the loop also enables it when the user had explicitly disabled it before. This is a common occurrence when the loss is computed over a validation set.
|
I just had this bug causing validation to hang when used with DistributedDataParallel. DistributedDataParallel checks torch.is_grad_enabled() when determining whether to perform synchronization before forward pass. Geomloss turns grad_enabled on after the first iteration, causing threads to hang waiting for synchronization. |
jeanfeydy
added a commit
that referenced
this pull request
Jun 18, 2022
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Indiscriminately enabling autograd at the end of the loop also enables it when the user
had explicitly disabled it before. This is a common occurrence when the loss is computed
over a validation set.
In my application this would stop the training with an out of memory error because autograd on the validation data quickly exhausts memory while it is fine in the training loop (sequential data with validation on full sequences but training on subsequences).