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(manifests): grant the cache initializer access to its TrainJob
- Initializer environment loading overrides dataclass defaults with None
- numGPU returns a nondeterministic value when a container requests multiple GPU-like resources
- fix(framework): stop Flux appending the Runtime args to the training command
- Add unit tests for metrics server setup
- chore: add SetupServer unit tests
- KEP-3562: [OptimizationJob] Parallelize TrainJob creation using WorkQueue at scale
- feat(cache): install data cache namespace RBAC from the Helm chart
- fix(framework): do not allocate a worker when runLauncherAsNode is set and numNodes is 1
- [BUG]: IndexableMemTable underflows on an empty batch and fetch_partitions can return a non-overlapping batch
- Docs
- Go not yet supported