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(cache): reuse one Flight connection per worker in the data cache writer
- fix(framework): apply runtime numNodes to MPI worker nodes
- fix(initializer): match S3 ignore patterns against the relative path
- HuggingFace dataset initializer fails with IndexError for dataset ids without a namespace (hf://imdb)
- fix(initializer): accept HF dataset ids without a namespace
- fix(controller): apply groupKindConcurrency to the TrainJob controller
- fix(framework): use the JobSet network subdomain for the coordinator address
- test: cover namespace resolution fallback in cache initializer
- Discussion: Add a stable label to all Pods of a TrainJob
- Forward AWS_* environment variables from DataCacheInitializer to LeaderWorkerSet containers
- Docs
- Go not yet supported