Dev

.github/workflows/ci.yml

@@ -1,49 +1,106 @@
-name: CI/CD
+name: Build and Push Docker Image
 
 on:
   push:
-    branches: [ '*' ]
+    branches:
+      - '**'
+    tags: ["v*.*.*"]
   pull_request:
-    branches: [ '*' ]
+    branches:
+      - '**'
+
+env:
+  REGISTRY: ghcr.io
+  IMAGE_NAME: ${{ github.repository }}
+  CACHE_FROM: type=registry,ref=${{ github.repository }}:buildcache
+  CACHE_TO: type=registry,ref=${{ github.repository }}:buildcache,mode=max
 
 jobs:
-  docker:
+  free-disk-space:
     runs-on: ubuntu-latest
-    if: github.event_name == 'push'
+    steps:
+      - name: Free Disk Space (Ubuntu)
+        uses: jlumbroso/free-disk-space@main
+        with:
+          tool-cache: false
+          android: true
+          dotnet: true
+          haskell: true
+          large-packages: true
+          docker-images: true
+          swap-storage: true
+
+  build-and-push:
+    runs-on: ubuntu-latest
+    needs: free-disk-space  # Ensure this job runs after freeing disk space
     permissions:
       contents: read
       packages: write
-
+      id-token: write
+
     steps:
-    - uses: actions/checkout@v3
-
-    - name: Set up Docker Buildx
-      uses: docker/setup-buildx-action@v2
-
-    - name: Log in to GitHub Container Registry
-      uses: docker/login-action@v2
-      with:
-        registry: ghcr.io
-        username: ${{ github.actor }}
-        password: ${{ secrets.GITHUB_TOKEN }}
-
-    - name: Extract metadata (tags, labels) for Docker
-      id: meta
-      uses: docker/metadata-action@v4
-      with:
-        images: ghcr.io/${{ github.repository }}
-        tags: |
-          type=semver,pattern={{version}}
-          type=sha,format=long
-          type=ref,event=branch
-          type=raw,value=latest,enable={{is_default_branch}}
-
-    - name: Build and push
-      uses: docker/build-push-action@v4
-      with:
-        context: .
-        push: true
-        tags: ${{ steps.meta.outputs.tags }}
-        labels: ${{ steps.meta.outputs.labels }}
-        cache-from: type=gha
-        cache-to: type=gha,mode=max
+      - name: Checkout repository
+        uses: actions/checkout@v4
+        with:
+          submodules: recursive
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+        with:
+          version: latest
+          driver-opts: |
+            image=moby/buildkit:latest
+
+      - name: Log into registry ${{ env.REGISTRY }}
+        if: github.event_name != 'pull_request'
+        uses: docker/login-action@v3
+        with:
+          registry: ${{ env.REGISTRY }}
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      # Cache Python dependencies
+      - name: Cache pip packages
+        uses: actions/cache@v3
+        with:
+          path: ~/.cache/pip
+          key: ${{ runner.os }}-pip-${{ hashFiles('**/requirements.txt', '**/pyproject.toml') }}
+          restore-keys: |
+            ${{ runner.os }}-pip-
+
+      # Set image name based on branch
+      - name: Set image name and tags
+        id: meta
+        uses: docker/metadata-action@v5
+        with:
+          images: ${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}
+          tags: |
+            # For main branch
+            type=raw,value=latest,enable=${{ github.ref == 'refs/heads/main' }}
+            type=sha,format=short,prefix=,enable=${{ github.ref == 'refs/heads/main' }}
+            # For other branches
+            type=raw,value=dev-latest,enable=${{ github.ref != 'refs/heads/main' }}
+            type=sha,format=short,prefix=dev-,enable=${{ github.ref != 'refs/heads/main' }}
+            # For tags
+            type=ref,event=tag
+            type=semver,pattern={{version}}
+            type=semver,pattern={{major}}.{{minor}}
+            type=semver,pattern={{major}}
+
+      - name: Build and push Docker image
+        id: build-and-push
+        uses: docker/build-push-action@v5
+        with:
+          context: .
+          push: ${{ github.event_name != 'pull_request' }}
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}
+          cache-from: |
+            type=gha
+            type=registry,ref=${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}:buildcache
+          cache-to: |
+            type=gha,mode=max
+            type=registry,ref=${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}:buildcache,mode=max
+          platforms: linux/amd64
+          build-args: |
+            BUILDKIT_INLINE_CACHE=1

Dockerfile

@@ -1,60 +1,68 @@
-FROM pytorch/pytorch:2.6.0-cuda12.4-cudnn9-devel
+FROM pytorch/pytorch:2.1.0-cuda12.1-cudnn8-devel
 
-WORKDIR /app
+# Set Zonos working directory
+WORKDIR /app/zonos
 
-# Install system dependencies
-RUN apt-get update && apt-get install -y \
-    build-essential \
+# System packages
+RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
+    --mount=type=cache,target=/var/lib/apt,sharing=locked \
+    apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends \
+    ffmpeg \
     libsndfile1 \
-    espeak-ng \
-    curl \
     git \
+    espeak-ng \
+    && apt-get clean \
     && rm -rf /var/lib/apt/lists/*
 
-# Install uv
-RUN pip install -U uv
-
-# Copy application code and submodules
-COPY . .
-
-# Initialize and update submodules
-RUN git submodule update --init --recursive --remote
-
-# Install dependencies with optimizations
-RUN uv pip install --system --no-build-isolation -e .[compile] && \
-    # Install flash-attention and other optimizations
-    pip install --no-build-isolation \
-    flash-attn \
-    mamba-ssm \
-    causal-conv1d
-
-# Create a non-root user and setup directories
-RUN useradd -m -u 1000 appuser && \
-    mkdir -p /home/appuser/.cache/huggingface && \
-    chown -R appuser:appuser /home/appuser/.cache && \
-    mkdir -p /app/uploads && \
-    chown -R appuser:appuser /app
-
-USER appuser
-
-# Set environment variables
-ENV PORT=8000
-ENV WORKERS=4
-ENV MODEL_TYPE="Transformer"
-ENV MODEL_CACHE_DIR="/home/appuser/.cache/huggingface"
-ENV PYTHONUNBUFFERED=1
-# CUDA optimization settings
-ENV CUDA_LAUNCH_BLOCKING=0
+# Install uv package manager
+RUN --mount=type=cache,target=/root/.cache/pip \
+    pip3 install --no-cache-dir uv
+
+# Clone Zonos directly into working directory
+RUN git clone --depth 1 https://github.com/Zyphra/Zonos.git . \
+    && git submodule update --init --recursive
+
+# Copy dependency specs and application code
+COPY requirements.txt pyproject.toml ./
+COPY app/ app/
+
+# Install basic Python dependencies first
+RUN --mount=type=cache,target=/root/.cache/pip \
+    uv pip install --system -r requirements.txt -e .[compile]
+
+# Install Flash Attention with specific compiler flags
 ENV TORCH_CUDA_ARCH_LIST="7.0;7.5;8.0;8.6+PTX"
-ENV CUDA_HOME="/usr/local/cuda"
-ENV MAX_JOBS=4
+ENV FLASH_ATTENTION_FORCE_BUILD=1
+RUN --mount=type=cache,target=/root/.cache/pip \
+    uv pip install --system --no-build-isolation \
+    git+https://github.com/Dao-AILab/[email protected]
+
+# Install remaining ML dependencies
+RUN --mount=type=cache,target=/root/.cache/pip \
+    uv pip install --system --no-build-isolation \
+    mamba-ssm==2.2.4 \
+    causal-conv1d==1.5.0.post8
+
+RUN --mount=type=cache,target=/root/.cache/pip \
+    uv pip install --system \
+    kanjize>=1.5.0 \
+    inflect>=7.5.0 \
+    && rm -rf /root/.cache/pip/*
+
+RUN --mount=type=cache,target=/root/.cache/pip \
+    uv pip install --system \
+    phonemizer>=3.3.0 \
+    sudachidict-full>=20241021 \
+    sudachipy>=0.6.10 \
+    && rm -rf /root/.cache/pip/*
 
-# Expose the port
-EXPOSE $PORT
+# Copy application code last
+COPY app/ app/
 
-# Add healthcheck
-HEALTHCHECK --interval=30s --timeout=10s --start-period=40s --retries=3 \
-    CMD curl -f http://localhost:$PORT/health || exit 1
+# Environment variables
+ENV PYTHONPATH=/app:/app/zonos \
+    USE_GPU=true \
+    PYTHONUNBUFFERED=1
 
-# Run the application with Gunicorn
-CMD ["sh", "-c", "gunicorn main:app --workers $WORKERS --worker-class uvicorn.workers.UvicornWorker --bind 0.0.0.0:$PORT --timeout 300 --worker-tmp-dir /dev/shm"]
+# Run the application
+CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "8000"]