trainer
https://github.com/kubeflow/trainer
Go
Distributed AI Model Training and LLM Fine-Tuning on Kubernetes
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Go not yet supported1 Subscribers
Add a CodeTriage badge to trainer
Help out
- Issues
- JobSet builder discards Info-based Command/Args/Image, breaking the documented EnforceMLPolicyPlugin contract
- fix(controller): preserve runtime logger context
- feat(manifests): Support high availability for the controller manager in the Helm chart
- runtimePatches targeting a job or container the runtime doesn't define are never rejected, because the patch is merged before validation
- chore(test): add unit tests for EventHandlerRegistrars and TrainJobStatus
- fix(controller): parallelize OptimizationJob trial creation with WorkQueue
- fix(runtimes): divide GPU count by tasks when setting Flux -g flag
- Expose custom trainer metrics via the /metrics scrape endpoint
- Document the custom metrics and the scraping configuration
- fix(manifests): expose the manager TLS options in the Helm chart
- Docs
- Go not yet supported