Kubeflow Trainer v2.1 Release #2859
Description
Creating this issue to track Kubeflow Trainer v2.1 release. Currently, we target mid October to prepare the first release candidate.
Milestone for the v2.1 release: https://github.com/kubeflow/trainer/milestone/6
- KEP-2655: Kubeflow Data Cache for distributed training on Kubernetes #2655
- Support for ResourcesPerNode in DeepSpeed Training Job Containers #2650
- Add PodTemplateOverrides API into TrainJob #2784
- Support JAX Runtimes #2442
- [GSoC] Project 10: Support Volcano Scheduler in Kubeflow Trainer #2671
- feat(api): Sync TrainJob JobsStatus from JobSet ReplicatedJobsStatus #2802
- feat(runtimes): Add LoRA/QLoRA/DoRA support in LLM Trainer V2 #2832
Please let me know if I miss any pending large features we should complete.
cc @kubeflow/kubeflow-trainer-team @kubeflow/kubeflow-sdk-team @akshaychitneni @Doris-xm @mahdikhashan @Labreo @zren11 @kannon92 @zren11 @mimowo @rudeigerc @sceneryback @Monokaix