Merge branch 'main' into mixtral

nod-ai · Nov 23, 2024 · b01d5a5 · b01d5a5
2 parents dd29409 + eacbd9b
commit b01d5a5
Show file tree

Hide file tree

Showing 426 changed files with 37,706 additions and 6,079 deletions.
diff --git a/.github/workflows/build_packages.yml b/.github/workflows/build_packages.yml
@@ -0,0 +1,158 @@
+# Copyright 2024 Advanced Micro Devices, Inc.
+#
+# Licensed under the Apache License v2.0 with LLVM Exceptions.
+# See https://llvm.org/LICENSE.txt for license information.
+# SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+
+name: Build packages
+
+on:
+  workflow_dispatch:
+  schedule:
+    # Runs at 11:00 AM UTC, which is 3:00 AM PST (UTC-8)
+    - cron: '0 11 * * *'
+
+permissions:
+  contents: read
+
+jobs:
+  # Note: metadata generation could happen in a separate trigger/schedule
+  # workflow. For cross platform builds, it's useful to just generate the
+  # metadata on Linux and pass that to later jobs using artifacts.
+  setup_metadata:
+    if: ${{ github.repository_owner == 'nod-ai' || github.event_name != 'schedule' }}
+    runs-on: ubuntu-24.04
+    outputs:
+      shark_package_version: ${{ steps.version.outputs.shark_package_version }}
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+        with:
+          submodules: false
+      - name: Setup Python
+        uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b # v5.3.0
+        with:
+          python-version: 3.12
+          cache: "pip"
+      - name: Install Python packages
+        run: |
+          pip install packaging
+          pip freeze
+      - name: Generate release candidate versions
+        id: version_rc
+        run: |
+          version_suffix="$(printf 'rc%(%Y%m%d)T')"
+          echo "version_suffix=${version_suffix}" >> $GITHUB_ENV
+          sharktank_package_version=$(python3 build_tools/python_deploy/compute_local_version.py --version-suffix=${version_suffix} sharktank)
+          shortfin_package_version=$(python3 build_tools/python_deploy/compute_local_version.py --version-suffix=${version_suffix} shortfin)
+          sharkai_package_version=$(python3 build_tools/python_deploy/compute_common_version.py -rc --version-suffix=${version_suffix} --write-json)
+      - name: Upload version_local.json
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882 # v4.4.3
+        with:
+          name: version_local
+          path: |
+            sharktank/version_local.json
+            shortfin/version_local.json
+            shark-ai/version_local.json
+
+  build_packages:
+    name: "${{ matrix.package }} :: ${{ matrix.platform }} :: ${{ matrix.python-version }}"
+    runs-on: ${{ matrix.runs-on }}
+    permissions:
+      contents: write
+    needs: [setup_metadata]
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          # Ubuntu packages.
+          - runs-on: ubuntu-24.04
+            platform: linux-x86_64
+            package: shark-ai
+            python-version: cp311-cp311  # Ignored (generic wheel), set for workflow naming
+          - runs-on: ubuntu-24.04
+            platform: linux-x86_64
+            package: sharktank
+            python-version: cp311-cp311  # Ignored (generic wheel), set for workflow naming
+          - runs-on: ubuntu-24.04
+            platform: linux-x86_64
+            package: shortfin
+            python-version: cp310-cp310
+          - runs-on: ubuntu-24.04
+            platform: linux-x86_64
+            package: shortfin
+            python-version: cp311-cp311
+          - runs-on: ubuntu-24.04
+            platform: linux-x86_64
+            package: shortfin
+            python-version: cp312-cp312
+          - runs-on: ubuntu-24.04
+            platform: linux-x86_64
+            package: shortfin
+            python-version: cp313-cp313
+          - runs-on: ubuntu-24.04
+            platform: linux-x86_64
+            package: shortfin
+            python-version: cp313-cp313t
+
+          # TODO(#130): macOS platform
+          # TODO(#130): Windows platform
+
+    steps:
+    - name: Checkout repository
+      uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+      with:
+        path: "c" # Windows can hit path length limits, so use a short path.
+        submodules: false
+
+    - name: Download version_local.json
+      uses: actions/download-artifact@fa0a91b85d4f404e444e00e005971372dc801d16 # v4.1.8
+      with:
+        name: version_local
+        path: ./c/
+        merge-multiple: true
+
+    - name: Build shark-ai (Linux x86_64)
+      if: "matrix.package == 'shark-ai' && matrix.platform == 'linux-x86_64'"
+      env:
+        OUTPUT_DIR: "${{ github.workspace }}/bindist"
+      run: |
+        [ -e ./bindist/* ] && rm ./bindist/*
+        ./c/build_tools/python_deploy/write_requirements.py --version-suffix=${version_suffix}
+        ./c/shark-ai/build_tools/build_linux_package.sh
+
+    - name: Build sharktank (Linux x86_64)
+      if: "matrix.package == 'sharktank' && matrix.platform == 'linux-x86_64'"
+      env:
+        OUTPUT_DIR: "${{ github.workspace }}/bindist"
+      run: |
+        [ -e ./bindist/* ] && rm ./bindist/*
+        ./c/sharktank/build_tools/build_linux_package.sh
+
+    - name: Build shortfin (Linux x86_64, ${{ matrix.python-version }})
+      if: "matrix.package == 'shortfin' && matrix.platform == 'linux-x86_64'"
+      env:
+        OUTPUT_DIR: "${{ github.workspace }}/bindist"
+        OVERRIDE_PYTHON_VERSIONS: "${{ matrix.python-version }}"
+      run: |
+        [ -e ./bindist/* ] && rm ./bindist/*
+        ./c/shortfin/build_tools/build_linux_package.sh
+
+    - name: Upload python wheels
+      uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882 # v4.4.3
+      with:
+        if-no-files-found: error
+        name: snapshot-${{ matrix.package }}-${{ matrix.platform }}-${{ matrix.python-version }}
+        path: bindist
+
+    - name: Release python wheels
+      uses: ncipollo/release-action@2c591bcc8ecdcd2db72b97d6147f871fcd833ba5 # v1.14.0
+      with:
+        artifacts: bindist/*.whl
+        tag: "dev-wheels"
+        name: "dev-wheels"
+        body: "Automatic snapshot release of shark-ai python wheels."
+        removeArtifacts: false
+        allowUpdates: true
+        replacesArtifacts: true
+        makeLatest: false
diff --git a/.github/workflows/ci-llama-large-tests.yaml b/.github/workflows/ci-llama-large-tests.yaml
@@ -0,0 +1,93 @@
+# Copyright 2024 Advanced Micro Devices, Inc.
+#
+# Licensed under the Apache License v2.0 with LLVM Exceptions.
+# See https://llvm.org/LICENSE.txt for license information.
+# SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+
+name: Llama Benchmarking Tests
+
+on:
+  workflow_dispatch:
+  schedule:
+    # Weekdays at 4:00 AM UTC = 9:00 PM PST.
+    - cron: "0 4 * * 1-5"
+
+concurrency:
+  # A PR number if a pull request and otherwise the commit hash. This cancels
+  # queued and in-progress runs for the same PR (presubmit) or commit
+  # (postsubmit). The workflow name is prepended to avoid conflicts between
+  # different workflows.
+  group: ${{ github.workflow }}-${{ github.event.number || github.sha }}
+  cancel-in-progress: true
+
+jobs:
+  test_llama_large:
+    if: ${{ github.repository_owner == 'nod-ai' || github.event_name != 'schedule' }}
+    name: "Llama Benchmarking Tests"
+    strategy:
+      matrix:
+        version: [3.11]
+      fail-fast: false
+    runs-on: llama-mi300x-1
+    defaults:
+      run:
+        shell: bash
+    env:
+      PIP_CACHE_DIR: "${{ github.workspace }}/.pip-cache"
+      VENV_DIR: ${{ github.workspace }}/.venv
+    steps:
+      - name: Get Current Date
+        id: date
+        run: echo "::set-output name=date::$(date +'%Y-%m-%d')"
+
+      - name: "Setting up Python"
+        id: setup_python
+        uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b # v5.3.0
+        with:
+          python-version: ${{matrix.version}}
+
+      - name: "Checkout Code"
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+
+      - name: Cache Pip Packages
+        uses: actions/cache@6849a6489940f00c2f30c0fb92c6274307ccb58a # v4.1.2
+        id: cache-pip
+        with:
+          path: ${{ env.PIP_CACHE_DIR }}
+          key: pip-${{ steps.setup_python.outputs.python-version }}-${{ hashFiles('*requirements.txt') }}
+
+      - name: Install pip deps
+        run: |
+          python -m pip install --no-compile --upgrade pip
+          # Note: We install in three steps in order to satisfy requirements
+          # from non default locations first. Installing the PyTorch CPU
+          # wheels saves multiple minutes and a lot of bandwidth on runner setup.
+          pip install --no-compile -r pytorch-cpu-requirements.txt
+          pip install --no-compile -r requirements.txt -r sharktank/requirements-tests.txt -e sharktank/
+
+          # Install latest iree-turbine.
+          pip install --no-compile -f https://iree.dev/pip-release-links.html --src deps \
+            -e "git+https://github.com/iree-org/iree-turbine.git#egg=iree-turbine"
+
+
+          # Test with nightly releases, not what iree-turbine uses.
+          pip install -f https://iree.dev/pip-release-links.html --upgrade --pre \
+            iree-base-compiler \
+            iree-base-runtime
+
+      - name: Run llama tests
+        run: pytest sharktank/tests/models/llama/benchmark_amdgpu_test.py -v -s --run-nightly-llama-tests --iree-hip-target=gfx942 --html=out/llm/llama/benchmark/index.html
+
+      - name: Deploy to GitHub Pages
+        uses: peaceiris/actions-gh-pages@4f9cc6602d3f66b9c108549d475ec49e8ef4d45e # v4.0.0
+        with:
+          github_token: ${{ secrets.SHARK_PLATFORM_GH_TOKEN }}
+          publish_dir: ./out/llm/llama/benchmark
+          destination_dir: ./llm/llama/benchmark
+          keep_files: true
+
+      - name: Upload llama executable files
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882 # v4.4.3
+        with:
+          name: llama-files
+          path: ${{ github.workspace }}/${{ steps.date.outputs.date }}
diff --git a/.github/workflows/ci-llama-quick-tests.yaml b/.github/workflows/ci-llama-quick-tests.yaml
@@ -0,0 +1,85 @@
+# Copyright 2024 Advanced Micro Devices, Inc.
+#
+# Licensed under the Apache License v2.0 with LLVM Exceptions.
+# See https://llvm.org/LICENSE.txt for license information.
+# SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+
+name: Llama Benchmarking 8B Tests
+
+on:
+  workflow_dispatch:
+  pull_request:
+  push:
+    branches:
+      - main
+
+concurrency:
+  # A PR number if a pull request and otherwise the commit hash. This cancels
+  # queued and in-progress runs for the same PR (presubmit) or commit
+  # (postsubmit). The workflow name is prepended to avoid conflicts between
+  # different workflows.
+  group: ${{ github.workflow }}-${{ github.event.number || github.sha }}
+  cancel-in-progress: true
+
+jobs:
+  test_llama_quick:
+    name: "Llama Benchmarking 8B Tests"
+    strategy:
+      matrix:
+        version: [3.11]
+      fail-fast: false
+    runs-on: llama-mi300x-1
+    defaults:
+      run:
+        shell: bash
+    env:
+      PIP_CACHE_DIR: "${{ github.workspace }}/.pip-cache"
+      VENV_DIR: ${{ github.workspace }}/.venv
+    steps:
+      - name: Get Current Date
+        id: date
+        run: echo "::set-output name=date::$(date +'%Y-%m-%d')"
+
+      - name: "Setting up Python"
+        id: setup_python
+        uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b # v5.3.0
+        with:
+          python-version: ${{matrix.version}}
+
+      - name: "Checkout Code"
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+
+      - name: Cache Pip Packages
+        uses: actions/cache@6849a6489940f00c2f30c0fb92c6274307ccb58a # v4.1.2
+        id: cache-pip
+        with:
+          path: ${{ env.PIP_CACHE_DIR }}
+          key: pip-${{ steps.setup_python.outputs.python-version }}-${{ hashFiles('*requirements.txt') }}
+
+      - name: Install pip deps
+        run: |
+          python -m pip install --no-compile --upgrade pip
+          # Note: We install in three steps in order to satisfy requirements
+          # from non default locations first. Installing the PyTorch CPU
+          # wheels saves multiple minutes and a lot of bandwidth on runner setup.
+          pip install --no-compile -r pytorch-cpu-requirements.txt
+          pip install --no-compile -r requirements.txt -r sharktank/requirements-tests.txt -e sharktank/
+
+          # Install latest iree-turbine.
+          pip install --no-compile -f https://iree.dev/pip-release-links.html --src deps \
+            -e "git+https://github.com/iree-org/iree-turbine.git#egg=iree-turbine"
+
+
+          # Test with nightly releases, not what iree-turbine uses.
+          pip install -f https://iree.dev/pip-release-links.html --upgrade --pre \
+            iree-base-compiler \
+            iree-base-runtime
+
+      - name: Run llama 8b f16 decomposed test
+        run: pytest sharktank/tests/models/llama/benchmark_amdgpu_test.py -v -s --iree-hip-target=gfx942 --run-quick-llama-test
+
+      - name: Upload llama executable files
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882 # v4.4.3
+        with:
+          name: llama-files
+          path: ${{ github.workspace }}/${{ steps.date.outputs.date }}