Skip to content

[test] fix: temporarily disable qwen2.5 omni unit tests#2759

Merged
ko3n1g merged 1 commit intomainfrom
yuya/disable-omni-unit-test
Mar 11, 2026
Merged

[test] fix: temporarily disable qwen2.5 omni unit tests#2759
ko3n1g merged 1 commit intomainfrom
yuya/disable-omni-unit-test

Conversation

@yaoyu-33
Copy link
Copy Markdown
Contributor

@yaoyu-33 yaoyu-33 commented Mar 11, 2026

What does this PR do?

Temporarily disables the TestQwen25OmniModel test class with @pytest.mark.pleasefixme to unblock CI.

Why?

The tests introduced in #2634 instantiate full-size HF vision and audio encoders (Qwen2_5OmniVisionEncoder, Qwen2_5OmniAudioEncoder) at real Qwen2.5-Omni-7B dimensions. test_shared_embedding_or_output_weight constructs two such models in a single test with a 50s timeout, which is causing CI timeouts on other PRs (e.g. #2004).

Follow-up needed

Replace the full-size HF encoders with lightweight dummy stubs so the tests run fast and can be re-enabled. cc @yuekaizhang

Changelog

  • Mark TestQwen25OmniModel with @pytest.mark.pleasefixme to skip in CI

Made with Cursor

Summary by CodeRabbit

  • Tests
    • Updated test annotations to track pending improvements for enhanced reliability.

Mark TestQwen25OmniModel with @pytest.mark.pleasefixme to skip in CI.
The tests instantiate full-size HF vision/audio encoders which causes
timeout (>50s). Needs to be refactored to use lightweight dummy stubs.

Signed-off-by: yaoyu-33 <yaoyu.094@gmail.com>
Made-with: Cursor
@yaoyu-33
Copy link
Copy Markdown
Contributor Author

/ok to test 6d6807e

@copy-pr-bot
Copy link
Copy Markdown

copy-pr-bot bot commented Mar 11, 2026

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.

@yaoyu-33 yaoyu-33 added the docs-only With great power comes great responsibility. label Mar 11, 2026
@ko3n1g ko3n1g merged commit 35d2e00 into main Mar 11, 2026
23 checks passed
@ko3n1g ko3n1g deleted the yuya/disable-omni-unit-test branch March 11, 2026 19:14
@coderabbitai
Copy link
Copy Markdown
Contributor

coderabbitai bot commented Mar 11, 2026

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info
⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 9f1ddd99-04c5-47c7-a169-94af0b69ca15

📥 Commits

Reviewing files that changed from the base of the PR and between 508166c and 6d6807e.

📒 Files selected for processing (1)
  • tests/unit_tests/models/qwen_omni/modeling_qwen25_omni/test_omni_model.py

📝 Walkthrough

Walkthrough

A pytest decorator (@pytest.mark.pleasefixme) was added to the TestQwen25OmniModel test class to mark it as requiring future work. No functional logic changes; purely an annotation adjustment.

Changes

Cohort / File(s) Summary
Test Marker Annotation
tests/unit_tests/models/qwen_omni/modeling_qwen25_omni/test_omni_model.py
Added @pytest.mark.pleasefixme decorator to the TestQwen25OmniModel class to indicate a TODO for replacing full-size HF vision/audio encoders with dummy stubs to prevent timeout issues.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~2 minutes

✨ Finishing Touches
  • 📝 Generate docstrings (stacked PR)
  • 📝 Generate docstrings (commit on current branch)
🧪 Generate unit tests (beta)
  • Create PR with unit tests
  • Post copyable unit tests in a comment
  • Commit unit tests in branch yuya/disable-omni-unit-test

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands and usage tips.

copy-pr-bot bot pushed a commit that referenced this pull request Mar 19, 2026
Signed-off-by: yaoyu-33 <yaoyu.094@gmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

docs-only With great power comes great responsibility.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants