medical_inference/config.json

{
    "model_type": "llama"
}