UCSC-VLAA / MedVLThinker-7B-SFT_m23k

MedVLThinker-7B-SFT_m23k

Code: https://github.com/UCSC-VLAA/MedVLThinker Project Page: https://ucsc-vlaa.github.io/MedVLThinker/

Model Description

MedVLThinker-7B-SFT_m23k is a 7B parameter medical vision-language model based on Qwen2.5-VL. This model has been trained using supervised fine-tuning on Med23k dataset.

Model Details

• Base Model : Qwen/Qwen2.5-VL-7B-Instruct
• Model Size : 7B parameters
• Training Method : Supervised Fine-tuning
• Training Data : Med23k dataset

Usage

Check here for demo images: https://github.com/UCSC-VLAA/MedVLThinker?tab=readme-ov-file#demo

from transformers import Qwen2_5_VLForConditionalGeneration, AutoTokenizer, AutoProcessor
from qwen_vl_utils import process_vision_info
import torch

# Load the model
model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
    "UCSC-VLAA/MedVLThinker-7B-SFT_m23k",
    torch_dtype=torch.bfloat16,
    device_map="auto"
)
processor = AutoProcessor.from_pretrained("UCSC-VLAA/MedVLThinker-7B-SFT_m23k")

# Example usage
messages = [
    {
        "role": "system",
        "content": "You will solve a problem/request. You should provide your thoughts within <think> </think> tags before providing the answer. Write your final answer within <answer> </answer> tags.",
    },
    {
        "role": "user",
        "content": [
            {
                "type": "image",
                "image": "path/to/medical/image.jpg",
            },
            {"type": "text", "text": "What can you see in this medical image?"},
        ],
    }
]

# Preparation for inference
text = processor.apply_chat_template(
    messages, tokenize=False, add_generation_prompt=True
)
image_inputs, video_inputs = process_vision_info(messages)
inputs = processor(
    text=[text],
    images=image_inputs,
    videos=video_inputs,
    padding=True,
    return_tensors="pt",
)
inputs = inputs.to("cuda")

# Inference
generated_ids = model.generate(**inputs, max_new_tokens=2048, temperature=0.6, top_p=0.95, do_sample=True)
generated_ids_trimmed = [
    out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
]
output_text = processor.batch_decode(
    generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
)
print(output_text)

Citation

@article{medvlthinker2025,
  title={MedVLThinker: Simple Baselines for Multimodal Medical Reasoning},
  author={Your Team},
  journal={arXiv preprint},
  year={2025}
}

License

This model is released under the Apache 2.0 license.