Skip to content

Multi-gpu vllm

Multi-gpu vllm #346

name: CLI CUDA TensorRT-LLM Tests
on:
workflow_dispatch:
push:
branches:
- main
pull_request:
branches:
- main
types:
- opened
- reopened
- synchronize
- labeled
- unlabeled
concurrency:
cancel-in-progress: true
group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
jobs:
cli_cuda_tensorrt_llm_tests:
if: ${{
(github.event_name == 'push') ||
(github.event_name == 'workflow_dispatch') ||
contains( github.event.pull_request.labels.*.name, 'cli') ||
contains( github.event.pull_request.labels.*.name, 'cuda') ||
contains( github.event.pull_request.labels.*.name, 'tensorrt_llm') ||
contains( github.event.pull_request.labels.*.name, 'cli_cuda_tensorrt_llm')
}}
runs-on: [single-gpu, nvidia-gpu, a10, ci]
container:
image: huggingface/optimum-nvidia:latest
options: --gpus all --volume /mnt/cache/.cache/huggingface:/mnt/cache/
steps:
- name: Checkout
uses: actions/checkout@v4
- name: Install dependencies
run: |
pip install -e .[testing]
- name: Run tests
run: |
pytest -x -s -k "cli and cuda and tensorrt_llm"