diff --git a/configs/volcano-trainjob-integration.yaml b/configs/volcano-trainjob-integration.yaml index 77fb7a9..0b14d7b 100644 --- a/configs/volcano-trainjob-integration.yaml +++ b/configs/volcano-trainjob-integration.yaml @@ -310,6 +310,9 @@ spec: replicatedJobs: - name: node template: + metadata: + labels: + trainer.kubeflow.org/trainjob-ancestor-step: trainer spec: template: metadata: @@ -376,7 +379,7 @@ spec: trainer: image: nvcr.io/nvidia/pytorch:24.10-py3 command: - - python + - torchrun - /workspace/scripts/train_nccl.py numNodes: 2 resourcesPerNode: