fix(refit): enforce configured timeouts in legacy GRPO - #4019
Open
seonjinn wants to merge 8 commits into
Open
Conversation
Signed-off-by: seonjinn <sna@nvidia.com>
Signed-off-by: seonjinn <sna@nvidia.com>
Signed-off-by: seonjinn <sna@nvidia.com>
…timeout Signed-off-by: seonjinn <sna@nvidia.com>
|
Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually. Contributors can view more details about this message here. |
Signed-off-by: seonjinn <sna@nvidia.com>
seonjinn
marked this pull request as ready for review
September 10, 2026 08:07
|
Auto-sync is disabled for ready for review pull requests in this repository. Workflows must be run manually. Contributors can view more details about this message here. |
Signed-off-by: seonjinn <sna@nvidia.com>
Signed-off-by: seonjinn <sna@nvidia.com> # Conflicts: # nemo_rl/algorithms/grpo.py # nemo_rl/weight_sync/factory.py # tests/unit/algorithms/test_grpo.py # tests/unit/weight_sync/test_weight_synchronizer.py
Contributor
Author
|
/ok to test 99a3492 |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Legacy GRPO did not pass a configured refit deadline into its weight synchronizer or raw IPC and collective waits. A failed refit could therefore leave
run_grpo.pywaiting indefinitely even though worker-side watchdog support already existed.This change:
policy.generation.refit_timeout_sfor legacy vLLM runs;Leaving the setting null preserves current behavior. SingleController keeps using
async_rl.generation_fleet_health.refit_timeout_s.Test plan
pyrefly check