fix: supcon oom - #48
Merged
Merged
Conversation
…ted batch The DDP single-forward fix doubled the logit tensor [2B, seq, vocab] causing OOM on single GPU. Split MLM head computation into two halves so only [B, seq, vocab] logits are materialized at a time.
RobertaForMaskedLM.forward() always materializes [N, seq, vocab] logits. With the concatenated 2B batch from the DDP fix, this doubled peak memory (~13 GB logits on batch_size=64, seq=512) causing OOM on A100-80GB. Introduce SplitHeadWrapper that runs the encoder on the full batch but applies the LM head per-chunk, halving peak logit memory. DDP wraps the wrapper so gradient sync works correctly.
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
No description provided.