Fix RL Training - vLLM sampler TypeError - #5101
Open
chiajunglien wants to merge 1 commit into
Open
Conversation
chiajunglien
requested review from
A9isha,
NuojCheng,
RissyRan,
SurbhiJainUSC,
abhinavclemson,
aireenmei,
bvandermoon,
darisoy,
dipannita08,
gagika,
gobbleturk,
hengtaoguo,
huytransformer,
igorts-git,
jiangjy1982,
khatwanimohit,
richjames0,
shralex,
shuningjin,
vipannalla and
xibinliu
as code owners
September 2, 2026 07:58
There was a problem hiding this comment.
Code Review
This pull request updates the __init__ method in maxtext_vllm_rollout.py by removing the converter argument from the super().__init__ call and removing the self.converter = converter assignment. The reviewer notes that while removing converter from super().__init__ is correct, removing self.converter = converter will break weight conversion because the base class expects this attribute. A suggestion is provided to restore the assignment.
- Remove invalid `converter` kwarg from `MaxTextVllmSampler.__init__` .
chiajunglien
force-pushed
the
emma/fix-rl-0902
branch
from
September 2, 2026 08:05
3d799a9 to
baad737
Compare
Codecov Report❌ Patch coverage is
📢 Thoughts on this report? Let us know! |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Description
This PR resolves critical failures in RL training pipelines. It restores a previously dropped fix for the
MaxTextVllmSamplerinitialization.DAG Error Log
Root Cause
TypeError: Commit4521fc5accidentally overwrote the fix applied in5bf4b88due to a stale branch merge. It reintroduced the invalidconverter=converterkwarg into the basetunixVllmSampler.__init__, triggering an immediateTypeErrorduring RL graph initialization.Solution
Rollout Logic: Removed the invalid
converterargument fromsuper().__init__inMaxTextVllmSampler, restoring the logic without breaking the newly added Qwen3.5 routing.Tests
Llama3.1-70b RL: https://cloudlogging.app.goo.gl/i7cbeR8hzSyt38s36
Checklist
Before submitting this PR, please make sure (put X in square brackets):
gemini-reviewlabel.