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
- Ignored JobSet deletion error during TrainJob deadline cleanup causes orphaned resources
- fix(controller): propagate JobSet deletion error on deadline expiry
- KEP-4162: TrainJob failure policy for restarts and disruption handling
- Flux plugin splits trainer arguments that contain spaces
- TrainJob immutability rules can be bypassed by adding or removing trainer, initializer or activeDeadlineSeconds
- TorchTune incorrectly detects LoRA config from unrelated argument values
- fix(runtimes): apply MPI numNodes to the node Job instead of the launcher
- Quick survey: your experience as a PyTorch Ecosystem maintainer
- [PERF]: Data cache does quadratic work at startup and on every query
- feat(runtime): resolve the GPU count from DRA ResourceClaimTemplates
- Docs
- Go not yet supported