I have created a chatbot application (based on python 3.10.10, langchain_community==0.2.5 and ollama LLM model, Ollama Embeddings model). I run this application on my local computer (which does not have a GPU), it is working fine.

But now, I want to run the model on server (which does not have any dependencies installed prior) therefore, I want to use Docker Image. Since I am using the Ollama, i have to use separate image of it. Therefore, I have decided to use docker-compose.yml file, where I have mentioned the Service of Ollama and Service of Chatbot. Apart from this, I have another service written is ollama-models-pull to pull the models of Ollama for LLM model and for Embeddings model.

I have my docker-compose.yml file as below:

services:
  ollama:
    container_name: ollama_v2
    image: ollama/ollama:latest
    restart: unless-stopped
    volumes:
      - "./ollamadata:/root/.ollama"   # Persist data and models
    ports:
      - "11434:11434"                  # Expose the Ollama service on port 11434
    healthcheck:
      test: ["CMD", "ollama", "list"]  # Check if Ollama is responding
      interval: 10s                    # Run the health check every 10 seconds
      timeout: 30s                     # Timeout for the health check is 30 seconds
      retries: 5                       # Retry 5 times before marking it as unhealthy
      start_period: 10s                # Wait 10 seconds before starting health checks
    networks:
      - ollama_network

  ollama-models-pull:
    container_name: ollama-models-pull
    image: curlimages/curl:latest       # Use curl to send HTTP requests
    depends_on:
      ollama:
        condition: service_healthy      # Wait for Ollama to be healthy
    command: >
      curl -X POST host.docker.internal:11434/api/pull -d '{"name":"mistral:latest"}'
      curl -X POST host.docker.internal:11434/api/pull -d '{"name":"nomic-embed-text"}'
    networks:
      - ollama_network

  chatbot:
    container_name: chatbot_v2
    build:
      context: ./    # The directory where Dockerfile and code are located
      dockerfile: Dockerfile
    restart: unless-stopped
    environment:
      - API_URL=host.docker.internal:11434     # Chatbot will access the Ollama API
    ports:
      - "8501:8501"                   # Expose chatbot on port 8080 (or any other port)
    depends_on:
      ollama-models-pull:
        condition: service_completed_successfully  # Wait for model pull to complete
    networks:
      - ollama_network

networks:
  ollama_network:
    driver: bridge

Problem: When I execute the docker-compose.yml file, it initialized three services as mentioned above. But after few seconds, the service ollama-models-pull gets exited for some reason by showing this message ollama-models-pull exited with code 0 .

Could you please share your knowledge about how to run Ollama inside the docker container, Pull the models of Ollama inside docker container and use them?

My end goal is to run the chatbot application to the machine where no dependencies is installed like in sever.