SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

Fix AMD CI not showing GPU (#27555)

Browse Source 
fix

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
This commit is contained in:
Yih-Dar
2023-11-17 10:44:37 +01:00
committed by GitHub
parent fe3ce061c4
commit d903abfccc
1 changed files with 13 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
.github/workflows/self-push-amd.yml vendored
View File

@@ -38,14 +38,16 @@ jobs:
runs-on: [self-hosted, docker-gpu, amd-gpu, '${{ matrix.machine_type }}', '${{ inputs.gpu_flavor }}'] runs-on: [self-hosted, docker-gpu, amd-gpu, '${{ matrix.machine_type }}', '${{ inputs.gpu_flavor }}']
container: container:
image: huggingface/transformers-pytorch-amd-gpu-push-ci # <--- We test only for PyTorch for now image: huggingface/transformers-pytorch-amd-gpu-push-ci # <--- We test only for PyTorch for now
options: --device /dev/kfd --device /dev/dri --env HIP_VISIBLE_DEVICES --env ROCR_VISIBLE_DEVICES --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/ options: --device /dev/kfd --device /dev/dri --env ROCR_VISIBLE_DEVICES --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
steps: steps:
- name: ROCM-SMI - name: ROCM-SMI
run: | run: |
rocminfo | grep "Agent" -A 14 rocm-smi
- name: Show HIP environment - name: ROCM-INFO
run: |
rocminfo | grep "Agent" -A 14
- name: Show ROCR environment
run: | run: |
echo "HIP: $HIP_VISIBLE_DEVICES"
echo "ROCR: $ROCR_VISIBLE_DEVICES" echo "ROCR: $ROCR_VISIBLE_DEVICES"
setup_gpu: setup_gpu:
@@ -57,7 +59,7 @@ jobs:
runs-on: [self-hosted, docker-gpu, amd-gpu, '${{ matrix.machine_type }}', '${{ inputs.gpu_flavor }}'] runs-on: [self-hosted, docker-gpu, amd-gpu, '${{ matrix.machine_type }}', '${{ inputs.gpu_flavor }}']
container: container:
image: huggingface/transformers-pytorch-amd-gpu-push-ci # <--- We test only for PyTorch for now image: huggingface/transformers-pytorch-amd-gpu-push-ci # <--- We test only for PyTorch for now
options: --device /dev/kfd --device /dev/dri --env HIP_VISIBLE_DEVICES --env ROCR_VISIBLE_DEVICES --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/ options: --device /dev/kfd --device /dev/dri --env ROCR_VISIBLE_DEVICES --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
outputs: outputs:
matrix: ${{ steps.set-matrix.outputs.matrix }} matrix: ${{ steps.set-matrix.outputs.matrix }}
test_map: ${{ steps.set-matrix.outputs.test_map }} test_map: ${{ steps.set-matrix.outputs.test_map }}
@@ -155,7 +157,7 @@ jobs:
runs-on: [self-hosted, docker-gpu, amd-gpu, '${{ matrix.machine_type }}', '${{ inputs.gpu_flavor }}'] runs-on: [self-hosted, docker-gpu, amd-gpu, '${{ matrix.machine_type }}', '${{ inputs.gpu_flavor }}']
container: container:
image: huggingface/transformers-pytorch-amd-gpu-push-ci # <--- We test only for PyTorch for now image: huggingface/transformers-pytorch-amd-gpu-push-ci # <--- We test only for PyTorch for now
options: --device /dev/kfd --device /dev/dri --env HIP_VISIBLE_DEVICES --env ROCR_VISIBLE_DEVICES --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/ options: --device /dev/kfd --device /dev/dri --env ROCR_VISIBLE_DEVICES --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
steps: steps:
# Necessary to get the correct branch name and commit SHA for `workflow_run` event # Necessary to get the correct branch name and commit SHA for `workflow_run` event
# We also take into account the `push` event (we might want to test some changes in a branch) # We also take into account the `push` event (we might want to test some changes in a branch)
@@ -207,10 +209,12 @@ jobs:
- name: ROCM-SMI - name: ROCM-SMI
run: | run: |
rocminfo | grep "Agent" -A 14 rocm-smi
- name: Show HIP environment - name: ROCM-INFO
run: |
rocminfo | grep "Agent" -A 14
- name: Show ROCR environment
run: | run: |
echo "HIP: $HIP_VISIBLE_DEVICES"
echo "ROCR: $ROCR_VISIBLE_DEVICES" echo "ROCR: $ROCR_VISIBLE_DEVICES"
- name: Environment - name: Environment