Skip to content

[Trackio] support trackio gpu logging#43192

Closed
kashif wants to merge 6 commits into
huggingface:mainfrom
kashif:trackio-gpu
Closed

[Trackio] support trackio gpu logging#43192
kashif wants to merge 6 commits into
huggingface:mainfrom
kashif:trackio-gpu

Conversation

@kashif

@kashif kashif commented Jan 9, 2026

Copy link
Copy Markdown
Contributor

What does this PR do?

This pull request updates the integration with the trackio experiment tracking library to require version 0.14.0 or higher, improves GPU metrics logging, and updates documentation and installation instructions accordingly. The changes ensure that users have the correct trackio version and are informed about required extras for GPU monitoring. It also refines how GPU metrics are collected and logged, especially in distributed training scenarios.

Trackio version requirement and installation:

  • Enforced a minimum trackio version of 0.14.0 in TrackioCallback and updated error messaging to instruct users to upgrade if needed.
  • Updated documentation and code comments to specify trackio>=0.14.0 and recommend installing the [gpu] extra for GPU system metrics. [1] [2] [3] [4]

GPU metrics logging improvements:

  • Refactored GPU metrics collection in TrackioCallback to use trackio.gpu.collect_gpu_metrics, with fallback and clear warnings if GPU monitoring is unavailable or the [gpu] extra is not installed.
  • Improved handling of distributed training environments by gathering and logging GPU metrics from all ranks.

Miscellaneous improvements:

  • Set auto_log_gpu=False when initializing a Trackio run to avoid duplicate or unintended GPU logging.
  • Cleaned up unused imports in integration_utils.py.
  • Defined _MIN_TRACKIO_VERSION constant for clarity and maintainability.

Fixes # (issue)

Before submitting

  • This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
  • Did you read the contributor guideline,
    Pull Request section?
  • Was this discussed/approved via a Github issue or the forum? Please add a link
    to it if that's the case.
  • Did you make sure to update the documentation with your changes? Here are the
    documentation guidelines, and
    here are tips on formatting docstrings.
  • Did you write any new necessary tests?

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.

@kashif kashif requested a review from abidlabs January 9, 2026 11:38
@HuggingFaceDocBuilderDev

Copy link
Copy Markdown

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

@abidlabs abidlabs requested a review from qgallouedec January 13, 2026 19:15
Comment thread docs/source/en/tasks/mask_generation.md Outdated
Comment thread docs/source/en/tasks/training_vision_backbone.md Outdated

@abidlabs abidlabs left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Haven't tested end-to-end but code changes look good

kashif and others added 3 commits January 13, 2026 20:45
Co-authored-by: Abubakar Abid <islamrealm@gmail.com>
Co-authored-by: Abubakar Abid <islamrealm@gmail.com>
@github-actions

Copy link
Copy Markdown
Contributor

View the CircleCI Test Summary for this PR:

https://huggingface.co/spaces/transformers-community/circle-ci-viz?pr=43192&sha=0346ed

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants