---
license: apache-2.0
base_model:
  - Qwen/Qwen2.5-VL-7B-Instruct
pipeline_tag: image-text-to-text
library_name: transformers
tags:
  - medical
  - multimodal
  - vqa
  - visual-grounding
  - chain-of-thought
  - reinforcement-learning
  - grpo
  - qwen2_5_vl
language:
  - en
datasets:
  - ifms111/UniReason-Med-Data
---

# UniReason-Med

UniReason-Med is a medical multimodal model for grounded reasoning over 2D medical images and slice-serialized 3D volumes.

It studies whether grounded reasoning supervision from abundant 2D medical images can improve 3D medical VQA when both modalities share a common reasoning interface. A single checkpoint processes either a 2D image or a 3D volume serialized as ordered slices, generating interleaved textual reasoning and localized visual evidence through shared bounding-box syntax and region-token injection.

- **Base model:** [Qwen/Qwen2.5-VL-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-VL-7B-Instruct)
- **Training data:** [ifms111/UniReason-Med-Data](https://huggingface.co/datasets/ifms111/UniReason-Med-Data)
- **Modalities:** image + text -> text
- **License:** Apache-2.0

## Model Description

UniReason-Med is trained to interleave free-form reasoning with localized visual evidence. During reasoning, the model emits bounding boxes over the input image; the referenced region is cropped and re-injected as additional visual context for the next reasoning step. The same shared interface is applied to 2D images and to 3D volumes serialized as ordered slice sequences.

## Training

The model is built with supervised fine-tuning followed by GRPO reinforcement learning. RL uses answer-correctness and format rewards rather than ground-truth localization-overlap rewards such as IoU or Dice.

## Intended Use and Limitations

- **Intended use:** research on medical multimodal reasoning, visual grounding, and 2D-to-3D transfer.
- **Out of scope:** this is a research artifact and is not a medical device. It must not be used for clinical diagnosis, treatment decisions, or real patient care.
- **Limitations:** outputs may be incorrect, incomplete, or biased; predicted bounding boxes are reasoning aids, not validated localization.

## License

Released under the Apache License 2.0, consistent with the base model Qwen2.5-VL-7B-Instruct.