feat: add auditwheel for manylinux

2026-07-01 01:37:53 -04:00 · 2025-12-11 01:56:28 +09:00
parent 7f2f8c579e
commit 61052b550b
3 changed files with 83 additions and 16 deletions
@@ -65,7 +65,7 @@ jobs:
      - name: Install build dependencies
        run: |
          sudo apt install -y ninja-build clang time
-          pip install -U pip setuptools==75.8.0 wheel packaging psutil
+          pip install -U pip setuptools==75.8.0 wheel packaging psutil auditwheel

      - name: Build wheels
        id: build_wheels
@@ -75,12 +75,23 @@ jobs:
          wheel_name=$(basename $(ls flash-attention/dist/*.whl | head -n 1))
          echo "WHEEL_NAME=$wheel_name" >> $GITHUB_OUTPUT

+      - name: Apply auditwheel repair
+        run: |
+          auditwheel show flash-attention/dist/${{ steps.build_wheels.outputs.WHEEL_NAME }}
+          auditwheel repair flash-attention/dist/${{ steps.build_wheels.outputs.WHEEL_NAME }} -w flash-attention/dist_manylinux
+          wheel_name=$(basename $(ls flash-attention/dist_manylinux/*.whl | head -n 1))
+          echo "WHEEL_NAME_MANYLINUX=$wheel_name" >> $GITHUB_OUTPUT
+
      - name: Install Test
        run: |
          pip uninstall -y flash-attn > /dev/null 2>&1
          pip install --no-cache-dir flash-attention/dist/${{ steps.build_wheels.outputs.WHEEL_NAME }}
          python -c "import flash_attn; print(flash_attn.__version__)"

+          pip uninstall -y flash-attn > /dev/null 2>&1
+          pip install --no-cache-dir flash-attention/dist_manylinux/${{ steps.build_wheels.outputs.WHEEL_NAME_MANYLINUX }}
+          python -c "import flash_attn; print(flash_attn.__version__)"
+
      - name: Upload Release Asset
        if: ${{ inputs.is-upload }}
        env:
@@ -88,12 +99,18 @@ jobs:
        run: |
          tag_name=${{ github.ref_name }}
          wheel_path="flash-attention/dist/${{ steps.build_wheels.outputs.WHEEL_NAME }}"
+          wheel_path_manylinux="flash-attention/dist_manylinux/${{ steps.build_wheels.outputs.WHEEL_NAME_MANYLINUX }}"

          # Check if the file exists
          if [ ! -f "$wheel_path" ]; then
            echo "Error: Wheel file not found at $wheel_path"
            exit 1
          fi
+          if [ ! -f "$wheel_path_manylinux" ]; then
+            echo "Error: Wheel file not found at $wheel_path_manylinux"
+            exit 1
+          fi

          # Upload the release asset using GitHub CLI
          gh release upload "$tag_name" "$wheel_path" --clobber
+          gh release upload "$tag_name" "$wheel_path_manylinux" --clobber
@@ -230,7 +230,7 @@ jobs:
        shell: bash
        run: |
          sudo apt install -y ninja-build clang
-          pip install -U pip setuptools==75.8.0 wheel packaging psutil
+          pip install -U pip setuptools==75.8.0 wheel packaging psutil auditwheel

      - name: Build wheels
        timeout-minutes: 2160
@@ -242,6 +242,13 @@ jobs:
          wheel_name=$(basename $(ls flash-attention/dist/*.whl | head -n 1))
          echo "WHEEL_NAME=$wheel_name" >> $GITHUB_OUTPUT

+      - name: Apply auditwheel repair
+        run: |
+          auditwheel show flash-attention/dist/${{ steps.build_wheels.outputs.WHEEL_NAME }}
+          auditwheel repair flash-attention/dist/${{ steps.build_wheels.outputs.WHEEL_NAME }} -w flash-attention/dist_manylinux
+          wheel_name=$(basename $(ls flash-attention/dist_manylinux/*.whl | head -n 1))
+          echo "WHEEL_NAME_MANYLINUX=$wheel_name" >> $GITHUB_OUTPUT
+
      - name: Install Test
        shell: bash
        run: |
@@ -249,6 +256,10 @@ jobs:
          pip install --no-cache-dir flash-attention/dist/${{ steps.build_wheels.outputs.WHEEL_NAME }}
          python -c "import flash_attn; print(flash_attn.__version__)"

+          pip uninstall -y flash-attn > /dev/null 2>&1
+          pip install --no-cache-dir flash-attention/dist_manylinux/${{ steps.build_wheels.outputs.WHEEL_NAME_MANYLINUX }}
+          python -c "import flash_attn; print(flash_attn.__version__)"
+
      - name: Upload Release Asset
        if: ${{ inputs.is-upload }}
        shell: bash
@@ -257,15 +268,21 @@ jobs:
        run: |
          tag_name=${{ github.ref_name }}
          wheel_path="flash-attention/dist/${{ steps.build_wheels.outputs.WHEEL_NAME }}"
+          wheel_path_manylinux="flash-attention/dist_manylinux/${{ steps.build_wheels.outputs.WHEEL_NAME_MANYLINUX }}"

          # Check if the file exists
          if [ ! -f "$wheel_path" ]; then
            echo "Error: Wheel file not found at $wheel_path"
            exit 1
          fi
+          if [ ! -f "$wheel_path_manylinux" ]; then
+            echo "Error: Wheel file not found at $wheel_path_manylinux"
+            exit 1
+          fi

          # Upload the release asset using GitHub CLI
          gh release upload "$tag_name" "$wheel_path" --clobber
+          gh release upload "$tag_name" "$wheel_path_manylinux" --clobber

      - name: Clean up
        shell: bash
@@ -122,24 +122,23 @@ Please note that depending on the combination of versions, it may not be possibl
 In some version combinations, you cannot build wheels on GitHub-hosted runners due to job time limitations.
 To build the wheels for these versions, you can use self-hosted runners.

+#### Setup x86_64 Runner
+
+Clone the repository and navigate to the self-hosted-runner directory.
+
 ```bash
 git clone https://github.com/mjun0812/flash-attention-prebuild-wheels.git
 cd flash-attention-prebuild-wheels/self-hosted-runner
 ```

-Install qemu-user-static for ARM64 support.
-
-```bash
-sudo apt install qemu-user-static
-```
-
-Edit `env` and `env.arm` files to set the environment variables.
+Create the environment file from the template.

 ```bash
 cp env.template env
-cp env.template env.arm
 ```

+Edit the `env` file to set the environment variables.
+
 ```bash
 # Registry Token for GitHub Personal Access Token
 PERSONAL_ACCESS_TOKEN=[Github Personal Access Token]
@@ -150,7 +149,7 @@ REGISTRY_TOKEN=[Runner Registry Token]
 RUNNER_LABELS=Linux,self-hosted
 ```

-Edit compose.yml file if you use repository folked from this repository.
+Edit the `compose.yml` file if you use a repository forked from this repository.

 ```yaml
 services:
@@ -170,6 +169,40 @@ services:
      args:
        GH_RUNNER_VERSION: 2.329.0
        TARGET_ARCH: x64
+```
+
+Build and run the docker container.
+
+```bash
+# Build and run
+docker compose build runner
+docker compose up -d runner
+```
+
+#### (Optional) Setup ARM64 Runner
+
+If you also want to build wheels for ARM64 architecture, follow these additional steps.
+
+Install qemu-user-static for ARM64 support.
+
+```bash
+sudo apt install qemu-user-static
+```
+
+Create the environment file for ARM64 runner.
+
+```bash
+cp env.template env.arm
+```
+
+Edit the `env.arm` file with the same configuration as the `env` file.
+
+Add the ARM64 runner service to your `compose.yml` file.
+
+```yaml
+services:
+  runner:
+    # ... (existing x86_64 runner configuration)

  runner-arm:
    privileged: true
@@ -178,7 +211,7 @@ services:
      - .env.arm
    environment:
      REPOSITORY_URL: https://github.com/[OWNER]/[REPOSITORY]
-      RUNNER_NAME: self-hosted-runner
+      RUNNER_NAME: self-hosted-runner-arm
      RUNNER_GROUP: default
      TARGET_ARCH: arm64
    build:
@@ -190,12 +223,12 @@ services:
        PLATFORM: linux/arm64
 ```

-Then, build and run the docker container.
+Build and run the ARM64 runner container.

 ```bash
-# Build and run
-docker compose build
-docker compose up -d
+# Build and run both x86_64 and ARM64 runners
+docker compose build runner-arm
+docker compose up -d runner-arm
 ```

 ### Getting One-Time Registry Token for GitHub Actions Runner