- Deployment
- Docker Model Runner
Deployment
Docker Model Runner
Use Hugging Face's one-line Docker Model Runner path for the fastest container-first local start.
This is the smallest local deployment surface currently shown in the Hugging Face “Use this model” panel.
docker model run hf.co/CohereLabs/North-Mini-Code-1.0
Why use this path
Good for fast local evaluation when you want the fewest moving parts between the Hub and a running model.
When not to use it
If you need explicit tool-call parsers, interleaved reasoning controls, or a deeply configurable OpenAI-compatible server, use /deployment/vllm instead.
The Hugging Face deploy panel provides the one-line launch command, but much less North-specific detail than the vLLM path. Treat Docker Model Runner as a speed-first option rather than the most customizable one.