feat(rollout): add --rollout-shuffle to permute prompts per epoch - #238
Draft
Rockdu wants to merge 1 commit into
Draft
feat(rollout): add --rollout-shuffle to permute prompts per epoch#238Rockdu wants to merge 1 commit into
Rockdu wants to merge 1 commit into
Conversation
Rockdu
force-pushed
the
kangrui/rollout-shuffle
branch
from
September 7, 2026 06:01
17f9fad to
036f9f9
Compare
Rockdu
force-pushed
the
kangrui/rollout-shuffle
branch
from
September 7, 2026 06:13
036f9f9 to
869272d
Compare
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What
--rollout-shuffle(off by default):RolloutDataSourcecallsDataset.shuffle(epoch_id)at construction, at every epoch wrap, and afterload(), exactly where radixark/miles does it.Dataset.shufflealready existed inmiles/utils/diffusion_data.py(seeded by--rollout-seed + epoch_id) but the RL data source never invoked it, so prompts were consumed in jsonl file order.Validation
tests/fast/rollout/test_rollout_shuffle.py(stage-a-cpu): off keeps file order; on yields a deterministic permutation of every prompt; the epoch wrap re-permutes andload()resumes the same order.Checklist
pre-commitpasses on the touched files