SmolLM2 on Ollama

Min VRAM: 3 GB

Pull

ollama pull smollm2

Run

ollama run smollm2

API

After running, Ollama serves an OpenAI-compatible API at http://localhost:11434/v1.

Model card