fix: apply GROUP_BY_LABEL sampler for in-batch embedding losses - #644
Closed
tonycoder-hub wants to merge 1 commit into
Closed
fix: apply GROUP_BY_LABEL sampler for in-batch embedding losses#644tonycoder-hub wants to merge 1 commit into
tonycoder-hub wants to merge 1 commit into
Conversation
`train_embeddings` compared the instantiated loss module against the loss classes, so the condition never held and `BatchSamplers.GROUP_BY_LABEL` was never applied. Compare `args.loss` instead, and reset the batch sampler to the default for other losses so it does not leak between training calls. Co-authored-by: Tony Coder <407243179@qq.com>
Author
|
Closing as stale — opened on or before 2026-08-17 and still unmerged. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
train_embeddings compared an instantiated loss module against loss classes, so loss in (BatchAllTripletLoss, ..., SupConLoss) was always False and GROUP_BY_LABEL never applied. Compare args.loss (the class) and reset the sampler between train() calls. Tests: python -m pytest tests/test_trainer.py -k batch_sampler or non_default_loss -v --no-cov -> 9 passed. Distinct from #643/#620/#579/#627.