kubeflow / trainer / 33225905693
70%

DEFAULT BRANCH: master
Ran
Jobs 1
Files 43
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 69.431%. Remained the same
33225905693

push

github

web-flow 
feat(docs): document MPIJob migration to TrainJob (#3968)

* feat(docs): document MPIJob migration to TrainJob

Signed-off-by: hari-599 <harisdarsan@gmail.com>

* Update docs/operator-guides/migration.md

Co-authored-by: Andrey Velichkevich <andrey.velichkevich@gmail.com>
Signed-off-by: Hari Sudarsan <73232897+hari-599@users.noreply.github.com>

* fix(docs): address MPI migration review comments

Signed-off-by: hari-599 <harisdarsan@gmail.com>

* Update docs/operator-guides/migration.md

Co-authored-by: Andrey Velichkevich <andrey.velichkevich@gmail.com>
Signed-off-by: Hari Sudarsan <73232897+hari-599@users.noreply.github.com>

* Update docs/operator-guides/migration.md

Co-authored-by: Andrey Velichkevich <andrey.velichkevich@gmail.com>
Signed-off-by: Hari Sudarsan <73232897+hari-599@users.noreply.github.com>

* Update docs/operator-guides/migration.md

Co-authored-by: Andrey Velichkevich <andrey.velichkevich@gmail.com>
Signed-off-by: Hari Sudarsan <73232897+hari-599@users.noreply.github.com>

* Update docs/operator-guides/migration.md

Co-authored-by: Andrey Velichkevich <andrey.velichkevich@gmail.com>
Signed-off-by: Hari Sudarsan <73232897+hari-599@users.noreply.github.com>

* fix(docs): explain MPI process count configuration

Signed-off-by: hari-599 <harisdarsan@gmail.com>

* docs: address review feedback on MPI migration guide

- Mention MPI Operator alongside Training Operator v1 as the source APIs.
- Add a concrete MPI-enabled ClusterTrainingRuntime example so readers can
  see where the launcher/worker topology and MPI policy actually live, and
  reference it by name from the TrainJob instead of a placeholder.
- Add an interactive MPI example showing how to keep the launcher alive and
  run mpirun manually via kubectl exec.

Signed-off-by: Andrey Velichkevich <andrey.velichkevich@gmail.com>

* docs: note in-progress Intel MPI and MPICH backend support

Point readers at kubeflow/trainer#3712 so they can track or contribute to
extending mpiImpl... (continued)

2696 of 3883 relevant lines covered (69.43%)

0.81 hits per line

Jobs
ID Job ID Ran Files Coverage
1 33225905693.1 43
69.43
 GitHub Action Run
Source Files on build 33225905693