fix: batch_size must be multiple of num_generations, pad dataset#244
Merged
Conversation
…eeded TRL requires generation_batch_size % num_generations == 0. With batch_size=1 and num_generations=4, TRL rejects it. Fix: 1. Set per_device_train_batch_size = num_generations (minimum valid) 2. Pad dataset by repeating tasks if len(dataset) < batch_size With 1 task and num_generations=4: dataset padded to 4 rows, batch_size=4, generation_batch_size=4, 4 % 4 == 0 ✓ Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Supersedes #240. TRL requires
generation_batch_size % num_generations == 0. The previous fix (batch_size=1) violated this with num_generations=4.Fix:
per_device_train_batch_size = num_generations(minimum valid value)len(dataset) < batch_sizeWith client's config (1 task, num_generations=4):
Repeating tasks is fine for RL — same task with many rollouts = more learning signal per step.
Test plan
🤖 Generated with Claude Code