test_tensorrt_llm.yaml

name: TensorRT-LLM Tests

on:
  workflow_dispatch:
  push:
    branches: [main]
  pull_request:
    types: [opened, reopened, synchronize]

concurrency:
  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
  cancel-in-progress: true

jobs:
  pull_image_and_run_gpu_tests:
    runs-on: hf-dgx-01
    steps:
      - name: Checkout
        uses: actions/checkout@v3

      - name: Build image
        run: docker build
          --file docker/tensorrt_llm.dockerfile
          --build-arg USER_ID=$(id -u)
          --build-arg GROUP_ID=$(id -g)
          --tag opt-bench-tensorrt-llm:latest
          .

      - name: Run tests
        run: docker run
          --rm
          --net host
          --pid host
          --shm-size 64G
          --env USE_CUDA="1"
          --env USER_ID=$(id -u)
          --env GROUP_ID=$(id -g)
          --volume $(pwd):/workspace/optimum-benchmark
          --workdir /workspace/optimum-benchmark
          --gpus '"device=0,1"'
          --entrypoint /bin/bash
          opt-bench-tensorrt-llm:latest
          -c "pip install -e .[test] && pytest -k 'tensorrt_llm' -x"