compose.yaml

services:
  cpu-server:
    build:
      context: cpu-server/
    volumes:
      - transformers-cache:/.cache/huggingface/
      - torch-cache:/.cache/torch/
      - pip-cache:/.cache/pip/
      - dvc-cache:/var/task/.dvc/cache/
    environment:
      - TRANSFORMERS_CACHE=/.cache/huggingface/hub
      - TORCH_HOME=/.cache/torch/
      - XDG_CACHE_HOME=/.cache/
    links:
      - "gpu-server:gpuserver"
    ports:
      - "8001:8000"

  gpu-server:
    build:
      context: gpu-server/
    ports:
      - "8002:8000"
    volumes:
      - transformers-cache:/.cache/huggingface/
      - pip-cache:/.cache/pip/
      - dvc-cache:/var/task/.dvc/cache/
    environment:
      - TRANSFORMERS_CACHE=/.cache/huggingface/hub
      - XDG_CACHE_HOME=/.cache/
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]

  data-studio:
    build:
      context: data-studio/
    ports:
      - "8003:8000"

  mongo:
    image: mongo
    restart: always
    ports:
      - "27017:27017"

volumes:
  torch-cache:
    driver: local
    driver_opts:
      o: bind
      type: none
      device: ~/.cache/torch/
  transformers-cache:
    driver: local
    driver_opts:
      o: bind
      type: none
      device: ~/.cache/huggingface/
  pip-cache:
    driver: local
    driver_opts:
      o: bind
      type: none
      device: ~/.cache/pip/
  dvc-cache: