[debug] DebugUnderflowOverflow doesn't work with DP#12816
Merged
stas00 merged 1 commit intohuggingface:masterfrom Jul 21, 2021
Merged
[debug] DebugUnderflowOverflow doesn't work with DP#12816stas00 merged 1 commit intohuggingface:masterfrom
stas00 merged 1 commit intohuggingface:masterfrom
Conversation
4 tasks
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
As reported in #12815
DebugUnderflowOverflowbreaks under DP since the model gets new references to model sub-modules/params on replication and the old references are needed to track the model layer names.It might be possible to think of some workaround, most likely overriding
torch.nn.parallel.data_parallel.replicateto refresh the model references after the replication, but at the moment this is not required, since DDP works just fine. (or single GPU).So this PR adds a clean assert when DP is used, instead of a confusing exception. Update docs.
Fixes: #12815
@sgugger