Cache angle onnx

[cpaniaguam]

No description provided.

[coderabbitai]

Warning

Review limit reached

@cpaniaguam, you've reached your PR review limit, so we couldn't start this review.

Next review available in: 21 minutes

Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 64a4b88c-1ac7-4ba5-b005-8de73cf73e32

📥 Commits

Reviewing files that changed from the base of the PR and between ec0c6c1 and af938b7.

📒 Files selected for processing (2)

• .github/workflows/run_tests.yml
• tests/rl/test_rl_likelihood_builder.py

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch cache-angle-onnx

---

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

• X
• Mastodon
• Reddit
• LinkedIn

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

[Copilot]

Pull request overview

This PR aims to make RL ONNX-based likelihood tests more reliable in CI by avoiding on-demand downloads of the Angle ONNX model and reducing susceptibility to Hugging Face Hub rate limits.

Changes:

• Switched the RL likelihood builder test to load angle.onnx from a checked-in local fixture path instead of a filename that triggers Hugging Face download behavior.
• Added a GitHub Actions cache step intended to reuse the Hugging Face cache across CI runs.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
tests/rl/test_rl_likelihood_builder.py	Loads the Angle ONNX model from tests/fixtures/angle.onnx to keep tests deterministic and avoid network/HF rate limits.
.github/workflows/run_tests.yml	Adds caching for Hugging Face cache directories to reduce repeated downloads in CI.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[digicosmos86]

I think the test suite share the same issue. Theoretically, we use a huggingface function to download the model, which should cache the file automatically. However, in reality the file is downloaded every time a test is run, especially in slow tests. I think it's worth digging in a bit more

[digicosmos86]

I think this is a different caching mechanism here than what we want to achieve. This gets us caching across different runs, so that we don't have to download the onnx files again. However, since the onnx models are so small, I am not sure this is necessary. Can we spend some time looking into why the hf_download call is not cached during testing?

[cpaniaguam]

Sure, we can address this later.