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
- OptimizationJob accepts trainer env vars its trials can't use, then stalls without failing
- fix(api): reject TrainJob env names and resources that Pods cannot accept
- feat(api): implement MPI node groups (KEP-4064)
- Proposal: allow selecting ReplicatedJobs for PodGroupPolicy in sequential runtimes
- test(helm): add helm-unittest tests for the public ConfigMap
- chore(test): add helm-unittest cases for the public ConfigMap
- ExtractResourcePerNodeFromRuntime panics on a replicatedJob whose Job template omits metadata, so TrainJob creation is rejected with an opaque 500
- S3 initializer matches ignore_patterns against the full object key, so relative patterns never match
- fix(docs): use the step argument in the local execution log examples
- fix(initializer): handle Hugging Face dataset ids without a namespace
- Docs
- Go not yet supported