Stub the weight transfer in the AsyncGRPO epoch-stop test - #6826
Open
albertvillanova wants to merge 1 commit into
Open
Stub the weight transfer in the AsyncGRPO epoch-stop test#6826albertvillanova wants to merge 1 commit into
albertvillanova wants to merge 1 commit into
Conversation
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
This was referenced Aug 20, 2026
Open
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This PR fixes
TestEpochStop::test_epoch_stop_is_fork_independentfailing withImportError: vLLM >= 0.22.0 is required to use WeightTransferClient.Motivation
The test builds a real
AsyncGRPOTrainerbut only passesrollout_worker, leavingweight_transferunset. The trainer then constructs a realWeightTransferClient, which requires vLLM >= 0.22.0 and so cannot work in a test that has no vLLM server to talk to. Every other trainer-building test in the file already passes_StubWeightTransfer(); this one was missed.The failure is not visible today because the class is gated behind
is_ampere_or_newer()and the experimental CI runner is a T4, so the test is skipped. Moving that job to an L40S in #6741 makes it run for the first time.Changes
weight_transfer=_StubWeightTransfer()when building the trainer inTestEpochStop._trainNote
Cursor Bugbot is generating a summary for commit e19e807. Configure here.