Troubleshooting Ollama

Explore how to download, load, and use models with Ollama, both via Docker and Remote setups.

---

Ollama Inside Docker

🐳 Ollama Inside Docker

If Ollama is deployed inside Docker (e.g., using Docker Compose or Kubernetes), the service will be available:

• Inside the container: http://127.0.0.1:11434
• From the host: http://localhost:11435 (if exposed via host network)

Step 1: Check Available Models

docker exec -it openwebui curl http://ollama:11434/v1/models

From the host (if exposed):

curl http://localhost:11435/v1/models

Step 2: Download Llama 3.2

docker exec -it ollama ollama pull llama3.2

You can also download a higher-quality version (8-bit) from Hugging Face:

docker exec -it ollama ollama pull hf.co/bartowski/Llama-3.2-3B-Instruct-GGUF:Q8_0

BYO Ollama (External Ollama)

🛠️ Bring Your Own Ollama (BYO Ollama)

If Ollama is running on the host machine or another server on your network, follow these steps.

Step 1: Check Available Models

Local:

curl http://localhost:11434/v1/models

Remote:

curl http://<remote-ip>:11434/v1/models

Step 2: Set the OLLAMA_BASE_URL

export OLLAMA_HOST=<remote-ip>:11434

Step 3: Download Llama 3.2

ollama pull llama3.2

Or download the 8-bit version from Hugging Face:

ollama pull hf.co/bartowski/Llama-3.2-3B-Instruct-GGUF:Q8_0

---

You now have everything you need to download and run models with Ollama. Happy exploring!