update dependencies - liger + trl

[winglian]

Summary by CodeRabbit

• Chores
  • Updated package dependencies to use newer versions of liger-kernel and trl.
• New Features
  • Added an option to control importance sampling granularity in training settings.
• Tests
  • Enhanced test configuration to improve dataset processing performance.

[coderabbitai]

📝 Walkthrough

"""

Walkthrough

The requirements.txt file was updated to increase the version numbers of two dependencies: liger-kernel was upgraded from 0.6.0 to 0.6.1, and trl from 0.19.1 to 0.20.0. The test_activation_checkpointing_offload test was modified to add "dataset_processes": 4 to its configuration. The GRPOStrategy class was extended to include an importance_sampling_level argument from the trl config. The TRLConfig schema was updated to add an optional importance_sampling_level field.

Changes

Cohort / File(s)	Change Summary
Dependency version bumps requirements.txt	Updated liger-kernel from 0.6.0 to 0.6.1 and trl from 0.19.1 to 0.20.0. No other changes.
Test configuration update tests/e2e/patched/test_activation_checkpointing.py	Added "dataset_processes": 4 to the test configuration dictionary in test_activation_checkpointing_offload.
GRPOStrategy training args update src/axolotl/core/trainers/grpo/__init__.py	Extended set_training_args_kwargs method to add importance_sampling_level from the trl config if present.
TRLConfig schema update src/axolotl/utils/schemas/trl.py	Added optional importance_sampling_level field to TRLConfig model accepting "sequence", "token", or None with descriptive doc.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~15 minutes

Possibly related PRs

• upgrade liger to 0.6.0 #2893: Previously updated liger-kernel from 0.5.10 to 0.6.0 in requirements.txt.
• update trl to 0.18.2 #2814: Updated the version of the trl package in requirements.txt.

Suggested labels

ready to merge

Suggested reviewers

• SalmanMohammadi
• djsaunde
  """

Note

⚡️ Unit Test Generation is now available in beta!

Learn more here, or try it out under "Finishing Touches" below.

---

📜 Recent review details

Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 9d6f786 and 959319d.

📒 Files selected for processing (2)

• src/axolotl/core/trainers/grpo/__init__.py (1 hunks)
• src/axolotl/utils/schemas/trl.py (1 hunks)

🧰 Additional context used

🧠 Learnings (1)

src/axolotl/core/trainers/grpo/__init__.py (1)

Learnt from: NanoCode012
PR: #2952
File: src/axolotl/common/datasets.py:125-125
Timestamp: 2025-07-22T08:27:00.129Z
Learning: GRPO (cfg.rl == "grpo" or cfg.rl is RLType.GRPO) should be excluded from dataset label checking during preprocessing, while other RL methods should continue to have this functionality available.

⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (11)

• GitHub Check: PyTest from Source Dist (3.11, 2.7.0)
• GitHub Check: PyTest (3.11, 2.7.0)
• GitHub Check: PyTest from Source Dist (3.11, 2.6.0)
• GitHub Check: PyTest (3.11, 2.7.1)
• GitHub Check: PyTest from Source Dist (3.11, 2.7.1)
• GitHub Check: PyTest (3.11, 2.6.0)
• GitHub Check: pre-commit
• GitHub Check: test-axolotl-multigpu (126, 12.6.3, 3.11, 2.6.0, 2, true)
• GitHub Check: test-axolotl-multigpu (126, 12.6.3, 3.11, 2.7.1, vllm, 2, true)
• GitHub Check: test-axolotl-multigpu (126, 12.6.3, 3.11, 2.7.0, 2, true)
• GitHub Check: pre-commit

🔇 Additional comments (2)

src/axolotl/utils/schemas/trl.py (1)

83-89: LGTM! Well-structured configuration field addition.

The new importance_sampling_level field is properly implemented with:

• Correct type annotation using Literal for constrained string values
• Comprehensive description explaining the purpose and usage differences between GRPO and GSPO
• Appropriate default value of None for backward compatibility
• Consistent formatting with other fields in the class

src/axolotl/core/trainers/grpo/__init__.py (1)

85-88: LGTM! Proper integration of the new configuration field.

The implementation correctly:

• Uses a proper None check before adding the field to the kwargs dictionary
• Follows the established pattern used by other conditional field additions in the method
• Maintains consistency with the TRLConfig schema field naming

✨ Finishing Touches

• 📝 Generate Docstrings

🧪 Generate unit tests

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch updates

---

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

🪧 Tips

Chat

There are 3 ways to chat with CodeRabbit:

• Review comments: Directly reply to a review comment made by CodeRabbit. Example:
  • I pushed a fix in commit <commit_id>, please review it.
  • Explain this complex logic.
  • Open a follow-up GitHub issue for this discussion.
• Files and specific lines of code (under the "Files changed" tab): Tag @coderabbitai in a new review comment at the desired location with your query. Examples:
  • @coderabbitai explain this code block.
  • @coderabbitai modularize this function.
• PR comments: Tag @coderabbitai in a new PR comment to ask questions about the PR branch. For the best results, please provide a very specific query, as very limited context is provided in this mode. Examples:
  • @coderabbitai gather interesting stats about this repository and render them as a table. Additionally, render a pie chart showing the language distribution in the codebase.
  • @coderabbitai read src/utils.ts and explain its main purpose.
  • @coderabbitai read the files in the src/scheduler package and generate a class diagram using mermaid and a README in the markdown format.
  • @coderabbitai help me debug CodeRabbit configuration file.

Support

Need help? Create a ticket on our support page for assistance with any issues or questions.

Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments.

CodeRabbit Commands (Invoked using PR comments)

• @coderabbitai pause to pause the reviews on a PR.
• @coderabbitai resume to resume the paused reviews.
• @coderabbitai review to trigger an incremental review. This is useful when automatic reviews are disabled for the repository.
• @coderabbitai full review to do a full review from scratch and review all the files again.
• @coderabbitai summary to regenerate the summary of the PR.
• @coderabbitai generate docstrings to generate docstrings for this PR.
• @coderabbitai generate sequence diagram to generate a sequence diagram of the changes in this PR.
• @coderabbitai generate unit tests to generate unit tests for this PR.
• @coderabbitai resolve resolve all the CodeRabbit review comments.
• @coderabbitai configuration to show the current CodeRabbit configuration for the repository.
• @coderabbitai help to get help.

Other keywords and placeholders

• Add @coderabbitai ignore anywhere in the PR description to prevent this PR from being reviewed.
• Add @coderabbitai summary to generate the high-level summary at a specific location in the PR description.
• Add @coderabbitai anywhere in the PR title to generate the title automatically.

Documentation and Community

• Visit our Documentation for detailed information on how to use CodeRabbit.
• Join our Discord Community to get help, request features, and share feedback.
• Follow us on X/Twitter for updates and announcements.

[codecov]

Codecov Report

❌ Patch coverage is 66.66667% with 1 line in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/axolotl/core/trainers/grpo/__init__.py	50.00%	1 Missing ⚠️

📢 Thoughts on this report? Let us know!