fix(partial-rollout): cap max_new_tokens by prior response length#2122
Open
none0663 wants to merge 2 commits into
Open
fix(partial-rollout): cap max_new_tokens by prior response length#2122none0663 wants to merge 2 commits into
none0663 wants to merge 2 commits into
Conversation
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
In partial rollout, an aborted sample is resubmitted with its previously
generated tokens already attached (
sample.response_length > 0). However,sampling_params["max_new_tokens"]was still set to the fullrollout_max_response_len, allowing the engine to generate another fullbudget of tokens. As a result, the total response length could exceed
rollout_max_response_len(nothing downstream clamps it).This subtracts the already-generated length so the cumulative response stays
within
rollout_max_response_len. When the budget is already exhausted,max_new_tokensbecomes 0 and the sample is markedTRUNCATEDby theexisting guard.
Changes
rollout/sglang_rollout.py(generate)rollout/sglang_streaming_rollout.py(generate_streaming)Only affects runs with
--partial-rolloutand samples that already haveresponse tokens; fresh samples are unchanged.