Kokoro-FastAPI/docker/cpu/docker-compose.yml

name: kokoro-tts
services:
  kokoro-tts:
    # image: ghcr.io/remsky/kokoro-fastapi-cpu:latest
    # Uncomment below (and comment out above) to build from source instead of using the released image
    build:
      context: ../..
      dockerfile: docker/cpu/Dockerfile
    volumes:
      - ../../api/src:/app/api/src
      - ../../Kokoro-82M:/app/Kokoro-82M
    ports:
      - "8880:8880"
    environment:
      - PYTHONPATH=/app:/app/Kokoro-82M
      # ONNX Optimization Settings for vectorized operations
      - ONNX_NUM_THREADS=8  # Maximize core usage for vectorized ops
      - ONNX_INTER_OP_THREADS=4  # Higher inter-op for parallel matrix operations
      - ONNX_EXECUTION_MODE=parallel
      - ONNX_OPTIMIZATION_LEVEL=all
      - ONNX_MEMORY_PATTERN=true
      - ONNX_ARENA_EXTEND_STRATEGY=kNextPowerOfTwo
      
  # Gradio UI service [Comment out everything below if you don't need it]
  gradio-ui:
    # image: ghcr.io/remsky/kokoro-fastapi:latest-ui
    # Uncomment below (and comment out above) to build from source instead of using the released image
    build:
      context: ../../ui
    ports:
      - "7860:7860"
    volumes:
      - ../../ui/data:/app/ui/data
      - ../../ui/app.py:/app/app.py  # Mount app.py for hot reload
    environment:
      - GRADIO_WATCH=True  # Enable hot reloading
      - PYTHONUNBUFFERED=1  # Ensure Python output is not buffered