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
- fix(runtime): count each Pod once in the Workload-Aware Scheduling gang size
- KEP-3311: Support shared initializer for train jobs
- docs: add Torch distributed plugin configuration examples for FSDP and DeepSpeed in TrainingRuntime
- KEP-3562: Store trial history natively in `OptimizationJob` Status
- KEP-3562: [OptimizationJob] Add Normal and LogNormal distribution support to SearchSpace API
- KEP-3562: [OptimizationJob] Implement Secure Auth between Controller and gRPC Suggestion Service
- feat: Add .agents skill for add-python-sdk-model
- Cache initializer silently succeeds on Kubernetes API failures
- feat(api): add Normal and LogNormal distributions to SearchSpace
- fix: propagate TrainJob env vars to MPI launcher and worker containers
- Docs
- Go not yet supported