MastersThesis/src/paddle_ocr/docker-compose.workers.yml

# docker-compose.workers.yml - Multiple PaddleOCR workers for parallel Ray Tune
#
# Usage:
#   GPU (4 workers sharing GPU):
#     docker compose -f docker-compose.workers.yml up
#
#   CPU (4 workers):
#     docker compose -f docker-compose.workers.yml --profile cpu up
#
#   Scale workers (e.g., 8 workers):
#     NUM_WORKERS=8 docker compose -f docker-compose.workers.yml up
#
# Each worker runs on a separate port: 8001, 8002, 8003, 8004, ...

x-ocr-gpu-common: &ocr-gpu-common
  image: seryus.ddns.net/unir/paddle-ocr-gpu:latest
  volumes:
    - ../dataset:/app/dataset:ro
    - ../debugset:/app/debugset:rw
    - paddlex-cache:/root/.paddlex
  environment:
    - PYTHONUNBUFFERED=1
    - CUDA_VISIBLE_DEVICES=0
  deploy:
    resources:
      reservations:
        devices:
          - driver: nvidia
            count: 1
            capabilities: [gpu]
  restart: unless-stopped
  healthcheck:
    test: ["CMD", "python", "-c", "import urllib.request; urllib.request.urlopen('http://localhost:8000/health')"]
    interval: 30s
    timeout: 10s
    retries: 3
    start_period: 120s

x-ocr-cpu-common: &ocr-cpu-common
  image: seryus.ddns.net/unir/paddle-ocr-cpu:latest
  volumes:
    - ../dataset:/app/dataset:ro
    - ../debugset:/app/debugset:rw
    - paddlex-cache:/root/.paddlex
  environment:
    - PYTHONUNBUFFERED=1
  restart: unless-stopped
  healthcheck:
    test: ["CMD", "python", "-c", "import urllib.request; urllib.request.urlopen('http://localhost:8000/health')"]
    interval: 30s
    timeout: 10s
    retries: 3
    start_period: 120s

services:
  # GPU Workers (gpu profile) - share single GPU
  ocr-worker-1:
    <<: *ocr-gpu-common
    container_name: paddle-ocr-worker-1
    ports:
      - "8001:8000"
    profiles:
      - gpu

  ocr-worker-2:
    <<: *ocr-gpu-common
    container_name: paddle-ocr-worker-2
    ports:
      - "8002:8000"
    profiles:
      - gpu

  # CPU Workers (cpu profile) - for systems without GPU
  ocr-cpu-worker-1:
    <<: *ocr-cpu-common
    container_name: paddle-ocr-cpu-worker-1
    ports:
      - "8001:8000"
    profiles:
      - cpu

  ocr-cpu-worker-2:
    <<: *ocr-cpu-common
    container_name: paddle-ocr-cpu-worker-2
    ports:
      - "8002:8000"
    profiles:
      - cpu

volumes:
  paddlex-cache:
    name: paddlex-model-cache