NV-Reason-CXR

Chest X-Ray Clinical VLM

Non-Commercial

NV-Reason-CXR-3B is a vision-language model for chest X-ray interpretation, fine-tuned from Qwen2.5-VL-3B with SFT and GRPO. Given a frontal chest X-ray and a text prompt, it returns step-by-step reasoning followed by a concise answer, and it handles follow-up questions. It is intended for research and education, with outputs reviewed by qualified professionals.

X-Ray HuggingFace GitHub Paper Demo

Agent skills

Run it with an agent

Install the skill, then just ask your agent in plain language. medagent.nv_reason_cxr

install skill

npx skills add NVIDIA-Medtech/medical-AI-skills --skill nv-reason-cxr

Try saying

“Run NV-Reason-CXR on the synthetic fixture JSON and keep it offline for a smoke test.”

“Run NV-Reason-CXR on chest_xray.png and return the structured reasoning JSON for an engineering smoke test.”

“Run NV-Reason-CXR on samples/cxr_001.jpg and show the model's chain-of-thought response text and runtime identity.”

Skill on GitHub → Validation: Benchmark·Evals·Tests·Validators How agent skills work

Engineering & research use only — not a medical device, not for diagnosis or clinical decision-making. Skills follow the open Agent Skills spec.