Skip to content

fix(partial-rollout): cap max_new_tokens by prior response length #2122

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
none0663 wants to merge 2 commits into
base: main
Choose a base branch
from
Open

fix(partial-rollout): cap max_new_tokens by prior response length #2122

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
f2fc681
fix partial-rollout: cap max_new_tokens by prior response length
none0663 Jun 23, 2026
2888b79
Merge branch 'main' into fix-partial-rollout-rollout-max-response
none0663 Jun 23, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions slime/rollout/sglang_rollout.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -163,6 +163,9 @@ async def generate(args: Namespace, sample: Sample, sampling_params: dict[str, A
), f"Sample status is {sample.status}"

prompt_ids = _prepare_prompt_ids(sample, state.tokenizer, state.processor)
# partial-rollout reuses the sample; subtract prior progress so total <= rollout_max_response_len
if args.partial_rollout and sample.response_length > 0:
sampling_params["max_new_tokens"] = max(0, args.rollout_max_response_len - sample.response_length)

assert (
sampling_params["max_new_tokens"] >= 0
Expand Down
3 changes: 3 additions & 0 deletions slime/rollout/sglang_streaming_rollout.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -58,6 +58,9 @@ async def generate_streaming(args: Namespace, sample: Sample, sampling_params: d
), f"Sample status is {sample.status}"

prompt_ids = _prepare_prompt_ids(sample, state.tokenizer, state.processor)
# partial-rollout reuses the sample; subtract prior progress so total <= rollout_max_response_len
if args.partial_rollout and sample.response_length > 0:
sampling_params["max_new_tokens"] = max(0, args.rollout_max_response_len - sample.response_length)

assert (
sampling_params["max_new_tokens"] >= 0
Expand Down
Loading