docker-compose.yml

version: "3"

# Usefull links to understand how this works
# NVIDIA Docker: GPU Server Application Deployment Made Easy: https://developer.nvidia.com/blog/nvidia-docker-gpu-server-application-deployment-made-easy/
# Tensorflow docker: https://www.tensorflow.org/install/docker
# In order to "runtime: nvidia work", is necessary an updated docker-compose v1.27.4
# To verify if you can use GPU: https://www.tensorflow.org/guide/gpu
# Extra: Learn how to do multi-stage builds: https://docs.docker.com/develop/develop-images/multistage-build/
# Extra: docker save: https://docs.docker.com/engine/reference/commandline/save/
# Extra: docker load: https://docs.docker.com/engine/reference/commandline/load/

services: # containers
  tensorflow-gpu:
    runtime: nvidia
    environment:
      - NVIDIA_VISIBLE_DEVICES=all
    
    build:
      # Changes the current folder for docker-compose
      context: .
      dockerfile: ./Dockerfile
      args: 
        - username=crispa
        - userid=1000
    
    container_name: tf-gpu
    image: tf-gpu
    # Restart policies
    restart: unless-stopped
    
    # Logging configuration for the service
    logging: 
      driver: json-file
      options: 
        max-size: 50m
    
    ports: 
      - "8888:8888"
      - "6006:6006"
    
    volumes:
      - ./handson-ml2/:/home/crispa/src/

    command: /opt/conda/envs/tf-gpu/bin/jupyter-lab --ip='0.0.0.0' --port=8888 --no-browser