10 Pull requests merged by 4 people

• feat(sdk): Generate external Kubernetes and JobSet models

  #2466 merged Mar 5, 2025

• chore(test): Upload artifacts from dir

  #2473 merged Mar 5, 2025

• Make MPIMLPolicySource optional fields as a pointer

  #2472 merged Mar 5, 2025

• Implement UTs for PlainML plugin

  #2469 merged Mar 5, 2025

• chore(test): Add E2E tests for Kubeflow Trainer

  #2470 merged Mar 5, 2025

• KEP-2170: Add Kubeflow Trainer Pipeline Framework Design

  #2439 merged Mar 3, 2025

• fix: fix typos in script comments.

  #2465 merged Mar 3, 2025

• fix(sdk): resolve errors in deserialization

  #2457 merged Mar 2, 2025

• Bump JobSet to v0.8.0

  #2463 merged Mar 1, 2025

• Replace Kueue PodRequests helper with core k/k one

  #2461 merged Feb 28, 2025

5 Pull requests opened by 3 people

• [feature] migrate images to ghcr

  #2455 opened Feb 27, 2025

• Add Initialized and ComponentsCreated conditions to TrainJob API

  #2464 opened Mar 1, 2025

• WIP: Implemente torch plugin UTs

  #2471 opened Mar 5, 2025

• Add MPIMLPolicySource CRD defaulters

  #2474 opened Mar 5, 2025

• Use large runner for building container image

  #2475 opened Mar 5, 2025

8 Issues closed by 1 person

• KEP-2170: Add E2E tests for Kubeflow Training V2

  #2213 closed Mar 5, 2025

• Upgrade jobset version to v0.8.0

  #2447 closed Mar 1, 2025

• KEP-2170: Migrate the container resource calculation mechanism to k/k library

  #2280 closed Feb 28, 2025

• Flaky test: [It] should delete redundant Pods

  #1848 closed Feb 28, 2025

• Flaky test: [It] should update TFJob with desired status

  #1820 closed Feb 28, 2025

• Flaky test: [It] should delete job when expired time is up

  #1821 closed Feb 28, 2025

• Flaky test: [It] should delete designated Pod

  #1844 closed Feb 28, 2025

• Flaky test: [It] should create missing Pods

  #1838 closed Feb 28, 2025

9 Issues opened by 4 people

• Decouple UTs between Framework and Plugins packages

  #2468 opened Mar 3, 2025

• Use JobSet DependsOn API for TrainJob

  #2467 opened Mar 3, 2025

• Explore `uv` project manager for Kubeflow Python SDK

  #2462 opened Feb 28, 2025

• Flaky Test: TestDatasetIntegration.test_dataset_download[HuggingFace - Public dataset-huggingface-test_case0]

  #2460 opened Feb 28, 2025

• KEP-2170: Revisit TrainJob Created condition status type

  #2459 opened Feb 28, 2025

• KEP-2170: Add Kubeflow Trainer Pipeline Framework Concept page to Documentation

  #2458 opened Feb 28, 2025

• Distributed training with mutliple pods, with multi-gpu in each pod

  #2456 opened Feb 28, 2025

• Managing Pod Lifecycle in Distributed Training with TFJob

  #2454 opened Feb 27, 2025

• Strategies for Deleting Successful Pods without Affecting Task Execution in TFJob

  #2453 opened Feb 27, 2025

26 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Add Helm chart for kubeflow trainer

  #2435 commented on Mar 5, 2025 • 55 new comments

• KEP-2401: Kubeflow LLM Trainer V2

  #2410 commented on Mar 5, 2025 • 37 new comments

• KEP-2170: Adding validation webhook for v2 trainjob

  #2307 commented on Feb 28, 2025 • 17 new comments

• fix restart policy bug in mpi job UpdateJobConditions

  #2344 commented on Mar 5, 2025 • 5 new comments

• Upgrade: K8s 1.32

  #2448 commented on Mar 4, 2025 • 4 new comments

• Add internal-cert-controller disable flag

  #2426 commented on Mar 3, 2025 • 0 new comments

• Add e2e tests for runtimes v2

  #2406 commented on Mar 2, 2025 • 0 new comments

• KEP-2170: Add the manifests overlay for Kubeflow Training V2

  #2382 commented on Mar 1, 2025 • 0 new comments

• Imporved the release process of training operator

  #2359 commented on Feb 26, 2025 • 0 new comments

• Cap `nproc_per_node` based on the CPU resources of the node for PyTorch TrainJob

  #2407 commented on Mar 5, 2025 • 0 new comments

• Don't overwrite PYTHONUNBUFFERED in PyTorchJob (and others)

  #1921 commented on Mar 5, 2025 • 0 new comments

• Add unit tests that cover the `pkg/apply` package

  #2452 commented on Mar 5, 2025 • 0 new comments

• KEP-2170: Kubeflow Trainer V2 API

  #2170 commented on Mar 5, 2025 • 0 new comments

• mpi job bug

  #2334 commented on Mar 5, 2025 • 0 new comments

• Add migration guide from Training Operator to Kubeflow Trainer V2

  #2412 commented on Mar 4, 2025 • 0 new comments

• Create Slurm runtime for model training using V2 APIs

  #2249 commented on Mar 4, 2025 • 0 new comments

• Export Models to Kubeflow Model Registry

  #2438 commented on Feb 28, 2025 • 0 new comments

• Support Volcano Scheduler in Kubeflow Trainer

  #2437 commented on Feb 28, 2025 • 0 new comments

• Enable GPU Testing for LLM Blueprints

  #2432 commented on Feb 28, 2025 • 0 new comments

• Support TensorFlow Runtime

  #2443 commented on Feb 28, 2025 • 0 new comments

• Support JAX Runtimes

  #2442 commented on Feb 28, 2025 • 0 new comments

• Support richer volcano scheduling

  #2182 commented on Feb 28, 2025 • 0 new comments

• pytorchjob didn't create worker pod ,seems hang

  #2327 commented on Feb 28, 2025 • 0 new comments

• How can I change the default MASTER_ADDR in Pytorchjob?

  #2331 commented on Feb 28, 2025 • 0 new comments

• "zero-trust" security / networking for training jobs

  #2341 commented on Feb 27, 2025 • 0 new comments

• Migrate images in Dockerhub to GHCR

  #2446 commented on Feb 27, 2025 • 0 new comments