Fix the gpt-oss example on huggingface_hub>=1.32: require trackio>=0.38.1

[h-guo18]

What does this PR do?

Type of change: Bug fix

Replaces the stale trackio<0.21 cap in examples/gpt-oss/requirements.txt with trackio>=0.38.1. This fixes the trtllm (gpt-oss) example test, which feeds the required example-pr-required-check.

Root cause. datasets 5.1.0 was released on 2026-10-05 and requires huggingface_hub>=1.31. Installing the example's requirements now upgrades hub in the TRT-LLM image from 1.16.1 to 1.33.0. trackio<0.21 resolves to 0.20.2, which imports CommitOperationAdd from huggingface_hub.hf_api, and hub 1.32.0 stopped exporting it there. trl imports trackio when it writes the model card at the first checkpoint, so sft.py crashes even with --report_to none:

ImportError: cannot import name 'CommitOperationAdd' from 'huggingface_hub.hf_api'

Main's nightly example run on 2026-10-06 fails the same way.

Why a floor instead of the cap. The cap came from #1390. At that time the example still supported transformers 4.x (hub < 1.0), and newer trackio required hub >= 1.0. Since #1746 the example requires transformers>=5.3, which itself requires huggingface_hub>=1.3, so the cap no longer protects anything. 0.38.1 is the first trackio release that imports CommitOperationAdd from the top-level package.

Why not pin datasets<5.1. #2666 does that for the tf_min unit env. Here it would only make CI pass, because the image ships hub 1.16.1. A fresh pip install -r examples/gpt-oss/requirements.txt still resolves hub 1.33.0 with trackio 0.20.2 and hits the same error.

Usage

No API change.

Testing

• Ran the CI job's steps (pip install -e ".[hf,dev-test]", the example requirements, then pytest tests/examples/gpt-oss) in nvcr.io/nvidia/tensorrt-llm/release:1.3.0rc22. It matches the CI image (rc20) on transformers, datasets, accelerate and peft.
  • Before: test_gpt_oss_sft_toy fails. It resolves datasets 5.1.0, hub 1.33.0 and trackio 0.20.2, the same as CI.
  • After: it passes with trackio 0.40.0. The other 2 tests are skipped without --run-release, as in CI.
• The example configs default to report_to: trackio. With that setting and trackio_space_id: null, sft.py trains, logs to the local trackio store, writes the model card and saves the model.
• Imports in clean venvs:
  • trackio 0.20.2 imports with hub 1.31.0 and fails with 1.32.0 and 1.33.0.
  • With hub 1.33.0, trackio 0.38.0 fails, while 0.38.1 and 0.40.0 import.

Before your PR is "Ready for review"

Make sure you read and follow Contributor guidelines and your commits are signed (git commit -s -S).

Make sure you read and follow the Security Best Practices (e.g. avoiding hardcoded trust_remote_code=True, torch.load(..., weights_only=False), pickle, etc.).

• Is this change backward compatible?: ✅
• If you copied code from any other sources or added a new PIP dependency, did you follow guidance in CONTRIBUTING.md: N/A (existing dependency, new version bound)
• Did you write any new necessary tests?: N/A (covered by the existing test_gpt_oss_sft_toy)
• Did you update Changelog?: N/A (example requirements only)
• Did you get Claude approval on this PR?: N/A

Additional Information

This is a companion to #2666, which fixes the tf_min unit job for the same datasets release. Open PRs pick up the fix once they merge main, for example #2589.

🤖 Generated with Claude Code

Summary by CodeRabbit

• Bug Fixes
  • Improved compatibility when setting up the GPT-OSS example in newer Hugging Face Hub environments. This helps avoid setup issues related to changes in the Hub’s available interfaces.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

• Configuration used: Repository: NVIDIA/Model-Optimizer/.coderabbit.yaml
• Review profile: CHILL
• Plan: Enterprise
• Run ID: b82cf9df-0d7e-4c25-9759-6e831ebec089

📥 Commits

Reviewing files that changed from the base of the PR and between 01e3a30 and c9147c6.

📒 Files selected for processing (1)

• examples/gpt-oss/requirements.txt

Included review availability: This review used your included allowance. Your plan provides up to 12 included reviews per hour; 11 remain after this review.

---

📝 Walkthrough

Walkthrough

The GPT-OSS example now requires trackio>=0.38.1 instead of trackio<0.21. A comment notes an import incompatibility between older Trackio versions and huggingface_hub>=1.32.

Changes

GPT-OSS dependency update

Layer / File(s)	Summary
Trackio version constraint examples/gpt-oss/requirements.txt	The requirement changes to trackio>=0.38.1. A comment notes the CommitOperationAdd import incompatibility with huggingface_hub>=1.32.

Priority: ➖ Normal

Estimated code review effort: 1 (Trivial) | ~3 minutes

Change: Bug fix

Merge Risk: ⚪ Minimal · up to c9147

No actionable merge-blocking risk is established for this requirement update; normal checks can proceed.

🚥 Pre-merge checks | ✅ 6

✅ Passed checks (6 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly summarizes the main change: updating the gpt-oss example to require a compatible Trackio version for newer huggingface_hub releases.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff. Analyzed 0 functions across 0…
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Security Anti-Patterns	✅ Passed	The PR changes only examples/gpt-oss/requirements.txt. It updates the existing trackio constraint from <0.21 to >=0.38.1; it adds no new PIP dependency and changes no Python code. The diff int…

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Commit to this branch
• Create a new PR

• Autopilot · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

_{Comment @coderabbitai help to get the list of available commands.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 69.19%. Comparing base (01e3a30) to head (c9147c6).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #2669   +/-   ##
=======================================
  Coverage   69.19%   69.19%           
=======================================
  Files         620      620           
  Lines       69654    69654           
=======================================
  Hits        48197    48197           
  Misses      21457    21457           

Flag	Coverage Δ
examples-gpt-oss	13.45% <ø> (-0.02%)	⬇️
examples-hf_ptq	22.87% <ø> (-0.04%)	⬇️
examples-llm_eval	17.41% <ø> (ø)
unit	58.63% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[kevalmorabia97]

@h-guo18 Please see if #2649 fixed this

[h-guo18]

fixed in #2649 , closing this PR