Fix for RL Test by wdykas · Pull Request #3148 · NVIDIA/Megatron-LM

wdykas · 2026-01-29T18:22:18Z

What does this PR do ?

Bringing back one of the RL tests with some changes to help make tests more deterministic.

⚠️ For major changes (either in lines of code or in its impact), please make sure to first share a design doc with the team. If you're unsure what's the best way to do so, contact the @mcore-oncall.

Contribution process

flowchart LR
    A[Pre-checks] --> B[PR Tests]
    subgraph Code Review/Approval
        C1[Expert Review] --> C2[Final Review]
    end
    B --> C1
    C2 --> D[Merge]

Pre-checks

I want this PR in a versioned release and have added the appropriate Milestone (e.g., Core 0.8)
I have added relevant unit tests
I have added relevant functional tests
I have added proper typing to my code Typing guidelines
I have added relevant documentation
I have run the autoformatter.sh on my PR

Code review

The following process is enforced via the CODEOWNERS file for changes into megatron/core. For changes outside of megatron/core, it is up to the PR author whether or not to tag the Final Reviewer team.

For MRs into `main` branch

Feel free to message or comment the @mcore-oncall to help accelerate your merge into main. The less complex your PR is, the faster it will be approved and merged!

(Step 1): Add PR label `Expert Review`

(Step 2): Collect the expert reviewers reviews

Attach the Expert Review label when your PR is ready for review.
GitHub auto-assigns expert reviewers based on your changes. They will get notified and pick up your PR soon.

⚠️ Only proceed to the next step once all reviewers have approved, merge-conflict are resolved and the CI is passing.
Final Review might get declined if these requirements are not fulfilled.

(Step 3): Final Review

Add Final Review label
GitHub auto-assigns final reviewers based on your changes. They will get notified and pick up your PR soon.

(Optional Step 4): Cherry-pick into release branch

If this PR also needs to be merged into core_r* release branches, after this PR has been merged, select Cherry-pick to open a new PR into the release branch.

For MRs into `dev` branch

The proposed review process for `dev` branch is under active discussion.

MRs are mergable after one approval by either eharper@nvidia.com or zijiey@nvidia.com.

Merging your PR

Any member of core-adlr and core-nemo will be able to merge your PR.

copy-pr-bot · 2026-01-29T18:22:22Z

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

dimapihtar · 2026-01-29T21:17:00Z

/ok to test 8f77f6c

jon-barker

this is great - thanks!

…sting

wdykas · 2026-01-30T15:51:25Z

/ok to test 27254b4

ko3n1g

One comment regarding the model config

ko3n1g · 2026-01-30T20:44:10Z

  - "mem-allocated-bytes"
  - "mem-max-allocated-bytes"
+
+THROUGHPUT_TEST_PARAMS:


Please don’t introduce new top-level attributes. If you need to feed metadata into your job, please use the pre-existing ENV mapping. It will export all of its key value pairs to the shell session

@jon-barker I think I got this from you I think. Do you think we need this?

Yes, the script that checks the metrics uses these to determine which metrics to compare and which iteration steps to use

I'm happy for them to be moved into ENV_VARS @ko3n1g but I can't do it today and someone needs to check all the tests using this get updated. I can do it when i get back from vacation on 2/9

Ill do it in this MR and make sure the tests pass on gitlab before review

wdykas · 2026-02-02T15:30:39Z

/ok to test 02a5bc3

ko3n1g

Thank you!

wdykas · 2026-02-02T20:35:14Z

/ok to test 02a5bc3

wdykas · 2026-02-03T00:09:25Z

/ok to test 02a5bc3

Co-authored-by: Dmytro Pykhtar <37850217+dimapihtar@users.noreply.github.com>

fixes

a2f957c

wdykas requested review from a team as code owners January 29, 2026 18:22

ko3n1g requested a review from a team January 29, 2026 18:22

dimapihtar added Expert Review [deprecated] Apply this label to indicate that your PR is ready for expert review. complexity: medium labels Jan 29, 2026

Merge branch 'main' into rl-testing

8f77f6c

copy-pr-bot Bot temporarily deployed to nemo-ci January 29, 2026 21:17 Inactive

ko3n1g added this to the Core 0.16 milestone Jan 29, 2026

copy-pr-bot Bot temporarily deployed to nemo-ci January 29, 2026 21:17 Inactive

copy-pr-bot Bot had a problem deploying to nemo-ci January 29, 2026 21:17 Failure

copy-pr-bot Bot temporarily deployed to test January 29, 2026 21:17 Inactive

tdene approved these changes Jan 29, 2026

View reviewed changes

jon-barker approved these changes Jan 29, 2026

View reviewed changes

wdykas added 2 commits January 29, 2026 16:16

fix determinsm

3ec0b1a

Merge branch 'rl-testing' of github.com:wdykas/Megatron-LM into rl-te…

9f01472

…sting

sidsingh-nvidia approved these changes Jan 30, 2026

View reviewed changes

kvareddy approved these changes Jan 30, 2026

View reviewed changes

merge main

27254b4

copy-pr-bot Bot temporarily deployed to nemo-ci January 30, 2026 15:51 Inactive

copy-pr-bot Bot had a problem deploying to nemo-ci January 30, 2026 15:51 Failure

copy-pr-bot Bot temporarily deployed to test January 30, 2026 15:52 Inactive

ko3n1g requested changes Jan 30, 2026

View reviewed changes

wdykas added 2 commits February 2, 2026 06:43

fix env var

7368cf4

merge main

02a5bc3

copy-pr-bot Bot temporarily deployed to nemo-ci February 2, 2026 15:30 Inactive

copy-pr-bot Bot had a problem deploying to nemo-ci February 2, 2026 15:31 Failure

copy-pr-bot Bot temporarily deployed to nemo-ci February 2, 2026 15:31 Inactive

copy-pr-bot Bot temporarily deployed to test February 2, 2026 15:31 Inactive

ko3n1g approved these changes Feb 2, 2026

View reviewed changes

wdykas enabled auto-merge February 2, 2026 18:27

shanmugamr1992 approved these changes Feb 2, 2026

View reviewed changes

copy-pr-bot Bot temporarily deployed to nemo-ci February 3, 2026 00:28 Inactive

wdykas added this pull request to the merge queue Feb 3, 2026

Merged via the queue into NVIDIA:main with commit c4bea0a Feb 3, 2026
47 of 50 checks passed

wdykas deleted the rl-testing branch February 3, 2026 00:33

daiyaanarfeen pushed a commit to daiyaanarfeen/Megatron-LM that referenced this pull request Feb 23, 2026

Fix for RL Test (NVIDIA#3148)

95780cb

Co-authored-by: Dmytro Pykhtar <37850217+dimapihtar@users.noreply.github.com>

yangbofun pushed a commit to xlm-research/Megatron-LM that referenced this pull request May 22, 2026

Fix for RL Test (NVIDIA#3148)

604f1a1

Co-authored-by: Dmytro Pykhtar <37850217+dimapihtar@users.noreply.github.com>

terminator123 pushed a commit to 021ai/Megatron-LM that referenced this pull request Aug 3, 2026

Fix for RL Test (NVIDIA#3148)

4d45704

Co-authored-by: Dmytro Pykhtar <37850217+dimapihtar@users.noreply.github.com>

Uh oh!

Conversation

wdykas commented Jan 29, 2026

What does this PR do ?

Contribution process

Pre-checks

Code review

(Step 1): Add PR label Expert Review

(Step 2): Collect the expert reviewers reviews

(Step 3): Final Review

(Optional Step 4): Cherry-pick into release branch

Merging your PR

Uh oh!

copy-pr-bot Bot commented Jan 29, 2026

Uh oh!

dimapihtar commented Jan 29, 2026

Uh oh!

jon-barker left a comment

Choose a reason for hiding this comment

Uh oh!

wdykas commented Jan 30, 2026

Uh oh!

ko3n1g left a comment

Choose a reason for hiding this comment

Uh oh!

ko3n1g Jan 30, 2026

Choose a reason for hiding this comment

Uh oh!

wdykas Jan 30, 2026

Choose a reason for hiding this comment

Uh oh!

jon-barker Jan 30, 2026

Choose a reason for hiding this comment

Uh oh!

jon-barker Jan 30, 2026

Choose a reason for hiding this comment

Uh oh!

wdykas Jan 30, 2026

Choose a reason for hiding this comment

Uh oh!

wdykas commented Feb 2, 2026

Uh oh!

ko3n1g left a comment

Choose a reason for hiding this comment

Uh oh!

wdykas commented Feb 2, 2026

Uh oh!

wdykas commented Feb 3, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

8 participants

(Step 1): Add PR label `Expert Review`