xla-gpu-tests/test_pytorch_xla_spmd.yaml at main · marijaEf/xla-gpu-tests · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
apiVersion: v1
kind: Pod
metadata:
  labels:
    eidf/user: <your-username>
    kueue.x-k8s.io/queue-name: <project-namespace>-user-queue
  generateName: pytorch-xla-imagenet-test-
spec:
  containers:
  - args:
    - -c
    - |
      cd /workspace
      git clone -b ddp-benchmark https://github.com/marijaEf/xla.git
      cd xla
      python3 test/spmd/test_train_spmd_imagenet.py --fake_data --sharding batch --batch_size 128 --epochs 2 --log_steps 20
    command:
    - /bin/bash
    env:
    - name: OMP_NUM_THREADS
      value: '1'
    - name: PJRT_DEVICE
      value: 'CUDA'
    - name: XLA_NUM_DEVICES
      value: '1'
    image: herefortheimage/pytorch-xla-2.7.1-mlperf-resnet50:latest
    name: mlperf-resnet
    resources:
      limits:
        cpu: 16
        memory: 64Gi
        nvidia.com/gpu: 1
      requests:
        cpu: 16
        memory: 32Gi
        nvidia.com/gpu: 1
    volumeMounts:
    - mountPath: /mnt/ceph_rbd
      name: volume
    - mountPath: /experiments
      name: logs
    - mountPath: /dev/shm
      name: devshm
    workingDir: /workspace/ML
  nodeSelector:
    nvidia.com/gpu.product: NVIDIA-A100-SXM4-80GB
  restartPolicy: Never
  volumes:
  - name: volume
    persistentVolumeClaim:
      claimName: imagenet-dataset
  - name: logs
    persistentVolumeClaim:
      claimName: resnet-logs
  - emptyDir:
      medium: Memory
    name: devshm