From b9b70c34f1ff83e728bb9644bf8bf6939e4510c3 Mon Sep 17 00:00:00 2001 From: Kevin DeShawn <126115026+KevDevSha@users.noreply.github.com> Date: Wed, 31 Jul 2024 14:49:12 -0500 Subject: [PATCH 1/5] test cpu --- .github/workflows/pr-cpu.yaml | 26 ++++++++++++++++---------- 1 file changed, 16 insertions(+), 10 deletions(-) diff --git a/.github/workflows/pr-cpu.yaml b/.github/workflows/pr-cpu.yaml index 2dd1c0edab..c43f4c29eb 100644 --- a/.github/workflows/pr-cpu.yaml +++ b/.github/workflows/pr-cpu.yaml @@ -1,3 +1,4 @@ + name: PR CPU tests on: push: @@ -15,23 +16,28 @@ concurrency: cancel-in-progress: ${{ github.ref != 'refs/heads/main' }} jobs: pytest-cpu: - uses: mosaicml/ci-testing/.github/workflows/pytest-cpu.yaml@v0.0.9 + name: ${{ matrix.name }} + runs-on: linux-ubuntu-latest strategy: matrix: include: - name: "cpu-2.3.1" + pip_deps: "[all-cpu]" container: mosaicml/pytorch:2.3.1_cpu-python3.11-ubuntu20.04 markers: "not gpu" pytest_command: "coverage run -m pytest" - name: ${{ matrix.name }} - if: github.repository_owner == 'mosaicml' - with: - container: ${{ matrix.container }} - name: ${{ matrix.name }} - pip_deps: "[all-cpu]" - pytest-command: ${{ matrix.pytest_command }} - pytest-markers: ${{ matrix.markers }} - safe_directory: llm-foundry + steps: + - name: Checkout code + uses: actions/checkout@v2 + - name: Run PR CPU Tests + uses: mosaicml/ci-testing/.github/actions/pytest-cpu@v0.1.0 + with: + name: ${{ matrix.name }} + container: ${{ matrix.container }} + pip_deps: ${{ matrix.pip_deps }} + pytest_command: ${{ matrix.pytest_command }} + pytest_markers: ${{ matrix.markers }} + safe_directory: llm-foundry coverage: uses: ./.github/workflows/coverage.yaml name: Coverage Results From 77f3c7dac2b8e2c99ee7eb6aca91d7ae90740bbd Mon Sep 17 00:00:00 2001 From: Kevin DeShawn <126115026+KevDevSha@users.noreply.github.com> Date: Wed, 31 Jul 2024 14:49:52 -0500 Subject: [PATCH 2/5] Update pr-cpu.yaml --- .github/workflows/pr-cpu.yaml | 1 - 1 file changed, 1 deletion(-) diff --git a/.github/workflows/pr-cpu.yaml b/.github/workflows/pr-cpu.yaml index c43f4c29eb..9285d1c262 100644 --- a/.github/workflows/pr-cpu.yaml +++ b/.github/workflows/pr-cpu.yaml @@ -1,4 +1,3 @@ - name: PR CPU tests on: push: From 4d32556364610432e64da96098ec0aa53adf383f Mon Sep 17 00:00:00 2001 From: Kevin DeShawn <126115026+KevDevSha@users.noreply.github.com> Date: Wed, 31 Jul 2024 15:57:06 -0500 Subject: [PATCH 3/5] Update pr-cpu.yaml --- .github/workflows/pr-cpu.yaml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/workflows/pr-cpu.yaml b/.github/workflows/pr-cpu.yaml index 9285d1c262..2c85719756 100644 --- a/.github/workflows/pr-cpu.yaml +++ b/.github/workflows/pr-cpu.yaml @@ -16,7 +16,7 @@ concurrency: jobs: pytest-cpu: name: ${{ matrix.name }} - runs-on: linux-ubuntu-latest + runs-on: ubuntu-latest strategy: matrix: include: From bbc79386be49ac8bf23870d71e9426967eb049b4 Mon Sep 17 00:00:00 2001 From: Kevin DeShawn <126115026+KevDevSha@users.noreply.github.com> Date: Wed, 31 Jul 2024 16:30:14 -0500 Subject: [PATCH 4/5] update gpu tests --- .github/workflows/pr-gpu.yaml | 108 +++++++++++++++++++--------------- 1 file changed, 59 insertions(+), 49 deletions(-) diff --git a/.github/workflows/pr-gpu.yaml b/.github/workflows/pr-gpu.yaml index c5638e403d..702c8f2067 100644 --- a/.github/workflows/pr-gpu.yaml +++ b/.github/workflows/pr-gpu.yaml @@ -4,17 +4,19 @@ on: branches: - main - release/* - pull_request_target: + pull_request: branches: - main - release/** workflow_dispatch: +# Cancel old runs when a new commit is pushed to the same branch if not on main or dev concurrency: group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }} cancel-in-progress: ${{ github.ref != 'refs/heads/main' }} jobs: pytest-gpu-1: - uses: mosaicml/ci-testing/.github/workflows/pytest-gpu.yaml@v0.0.9 + name: ${{ matrix.name }} + runs-on: linux-ubuntu-latest strategy: fail-fast: false matrix: @@ -22,24 +24,27 @@ jobs: - name: "gpu-2.3.1-1" container: mosaicml/llm-foundry:2.3.1_cu121-latest markers: "gpu" - pytest_command: "coverage run -m pytest" pip_deps: "[all]" - name: ${{ matrix.name }} - if: github.repository_owner == 'mosaicml' - with: - container: ${{ matrix.container }} - git_repo: mosaicml/llm-foundry - mcloud-timeout: 1800 - name: ${{ matrix.name }} - pip_deps: ${{ matrix.pip_deps }} - pytest-command: ${{ matrix.pytest_command }} - pytest-markers: ${{ matrix.markers }} - python-version: 3.9 - gpu_num: 1 - secrets: - mcloud-api-key: ${{ secrets.MCLOUD_API_KEY }} + pytest_command: "coverage run -m pytest" + ci_repo_gpu_test_ref: v0.1.0 + steps: + - name: Run PR GPU Tests + uses: mosaicml/ci-testing/.github/actions/pytest-gpu@v0.1.0 + with: + container: ${{ matrix.container }} + git_repo: databricks-mosaic/runtime-private-plugins + mcloud_timeout: 1800 + name: ${{ matrix.name }} + pip_deps: ${{ matrix.pip_deps }} + pytest_command: ${{ matrix.pytest_command }} + pytest_markers: ${{ matrix.markers }} + python_version: 3.9 + gpu_num: 1 + mcloud_api_key: ${{ secrets.MCLOUD_API_KEY }} + ci_repo_gpu_test_ref: ${{ matrix.ci_repo_gpu_test_ref }} pytest-gpu-2: - uses: mosaicml/ci-testing/.github/workflows/pytest-gpu.yaml@v0.0.9 + name: ${{ matrix.name }} + runs-on: linux-ubuntu-latest strategy: fail-fast: false matrix: @@ -47,24 +52,27 @@ jobs: - name: "gpu-2.3.1-2" container: mosaicml/llm-foundry:2.3.1_cu121-latest markers: "gpu" - pytest_command: "coverage run -m pytest" pip_deps: "[all]" - name: ${{ matrix.name }} - if: github.repository_owner == 'mosaicml' - with: - container: ${{ matrix.container }} - git_repo: mosaicml/llm-foundry - mcloud-timeout: 1800 - name: ${{ matrix.name }} - pip_deps: ${{ matrix.pip_deps }} - pytest-command: ${{ matrix.pytest_command }} - pytest-markers: ${{ matrix.markers }} - python-version: 3.9 - gpu_num: 2 - secrets: - mcloud-api-key: ${{ secrets.MCLOUD_API_KEY }} + pytest_command: "coverage run -m pytest" + ci_repo_gpu_test_ref: v0.1.0 + steps: + - name: Run PR GPU Tests + uses: mosaicml/ci-testing/.github/actions/pytest-gpu@v0.1.0 + with: + container: ${{ matrix.container }} + git_repo: databricks-mosaic/runtime-private-plugins + mcloud_timeout: 1800 + name: ${{ matrix.name }} + pip_deps: ${{ matrix.pip_deps }} + pytest_command: ${{ matrix.pytest_command }} + pytest_markers: ${{ matrix.markers }} + python_version: 3.9 + gpu_num: 2 + mcloud_api_key: ${{ secrets.MCLOUD_API_KEY }} + ci_repo_gpu_test_ref: ${{ matrix.ci_repo_gpu_test_ref }} pytest-gpu-4: - uses: mosaicml/ci-testing/.github/workflows/pytest-gpu.yaml@v0.0.9 + name: ${{ matrix.name }} + runs-on: linux-ubuntu-latest strategy: fail-fast: false matrix: @@ -72,19 +80,21 @@ jobs: - name: "gpu-2.3.1-4" container: mosaicml/llm-foundry:2.3.1_cu121-latest markers: "gpu" - pytest_command: "coverage run -m pytest" pip_deps: "[all]" - name: ${{ matrix.name }} - if: github.repository_owner == 'mosaicml' - with: - container: ${{ matrix.container }} - git_repo: mosaicml/llm-foundry - mcloud-timeout: 1800 - name: ${{ matrix.name }} - pip_deps: ${{ matrix.pip_deps }} - pytest-command: ${{ matrix.pytest_command }} - pytest-markers: ${{ matrix.markers }} - python-version: 3.9 - gpu_num: 4 - secrets: - mcloud-api-key: ${{ secrets.MCLOUD_API_KEY }} + pytest_command: "coverage run -m pytest" + ci_repo_gpu_test_ref: v0.1.0 + steps: + - name: Run PR GPU Tests + uses: mosaicml/ci-testing/.github/actions/pytest-gpu@v0.1.0 + with: + container: ${{ matrix.container }} + git_repo: databricks-mosaic/runtime-private-plugins + mcloud_timeout: 1800 + name: ${{ matrix.name }} + pip_deps: ${{ matrix.pip_deps }} + pytest_command: ${{ matrix.pytest_command }} + pytest_markers: ${{ matrix.markers }} + python_version: 3.9 + gpu_num: 4 + mcloud_api_key: ${{ secrets.MCLOUD_API_KEY }} + ci_repo_gpu_test_ref: ${{ matrix.ci_repo_gpu_test_ref }} From 6bf63536626b2fc6327e27c12942a58eaf99020b Mon Sep 17 00:00:00 2001 From: Kevin DeShawn <126115026+KevDevSha@users.noreply.github.com> Date: Wed, 31 Jul 2024 16:49:09 -0500 Subject: [PATCH 5/5] Update pr-gpu.yaml --- .github/workflows/pr-gpu.yaml | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/.github/workflows/pr-gpu.yaml b/.github/workflows/pr-gpu.yaml index 702c8f2067..04f30b5f9c 100644 --- a/.github/workflows/pr-gpu.yaml +++ b/.github/workflows/pr-gpu.yaml @@ -32,7 +32,7 @@ jobs: uses: mosaicml/ci-testing/.github/actions/pytest-gpu@v0.1.0 with: container: ${{ matrix.container }} - git_repo: databricks-mosaic/runtime-private-plugins + git_repo: mosaicml/llm-foundry mcloud_timeout: 1800 name: ${{ matrix.name }} pip_deps: ${{ matrix.pip_deps }} @@ -60,7 +60,7 @@ jobs: uses: mosaicml/ci-testing/.github/actions/pytest-gpu@v0.1.0 with: container: ${{ matrix.container }} - git_repo: databricks-mosaic/runtime-private-plugins + git_repo: mosaicml/llm-foundry mcloud_timeout: 1800 name: ${{ matrix.name }} pip_deps: ${{ matrix.pip_deps }} @@ -88,7 +88,7 @@ jobs: uses: mosaicml/ci-testing/.github/actions/pytest-gpu@v0.1.0 with: container: ${{ matrix.container }} - git_repo: databricks-mosaic/runtime-private-plugins + git_repo: mosaicml/llm-foundry mcloud_timeout: 1800 name: ${{ matrix.name }} pip_deps: ${{ matrix.pip_deps }}