kubeflow
/
trainer
mirror of https://github.com/kubeflow/trainer.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Add PodNetwork plugin to KEP-2170 Job Pipeline Framework description (#2578) 

Signed-off-by: Yuki Iwai <yuki.iwai.tz@gmail.com>
This commit is contained in:
Yuki Iwai 2025-03-31 07:35:30 +09:00 committed by GitHub
parent 4db032add8
commit 3781eda0e6
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 2 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
docs/proposals/2170-kubeflow-trainer-v2/README.md
View File

@ -1704,6 +1704,7 @@ On the other hand, the Internal APIs are not exposed and could not add any opera
to any kind of resources like PodSpec.
- `EnforceMLPolicy`: This configure MachineLearning framework specific parameters (e.x, specified in TrainingRuntime `.spec.mlPolicy`)
to any kind of resources like PodSpec.
- `PodNetwork`: This identifies Pod-to-Pod communication network endpoints for each Pod and stores them to `RuntimeInfo`.
- `ComponentBuilder`: This builds Kubernetes resources leveraging `RuntimeInfo` and `TrainJob`.
`RuntimeInfo` is abstracted objects extracted from runtimes like TrainingRuntime and ClusterTrainingRuntime.
- `PostExecution Phase`:

2
docs/proposals/2170-kubeflow-trainer-v2/TrainerPipelineFramework.drawio.svg
View File

File diff suppressed because one or more lines are too long
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 438 KiB

After

Width:  |  Height:  |  Size: 456 KiB