Ollama Docker Compose Setup 🚀

This Docker Compose configuration sets up two Ollama services using the ollama/ollama image.

Services 🧑‍💻

1. Main Ollama API Service (ollama-cpu) 🔌:

• Purpose: Runs the Ollama API server exposed on port 11434.
• Health Check: Includes a health check to ensure the service is running.
• Resources: Allocates 8.5GiB of memory.
• Restart Policy: Restarts unless stopped.

2. Model Initialization Service (ollama-pull-llama-cpu) 🏗️:

• Purpose: Pulls Llama models (like llama3.2, llama2, and nomic-embed-text) on container startup.
• Dependencies: Depends on the main Ollama service (ollama-cpu) to ensure models are downloaded before starting.
• Restart Policy: Does not restart after execution.

Configuration Details ⚙️:

• Volumes: Both services use a shared volume (ollama_storage) to store data.
• Networks: The services are connected via a custom network (demo).
• Profiles: Both services are configured for CPU usage.

This configuration is designed for setting up Ollama in a CPU environment, initializing models on startup, and serving them via the API.

Run the Services 💻

To start the services with the CPU profile, use the following command:

docker-compose --profile cpu up