Instructions to use lingshu-medical-mllm/Lingshu-32B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use lingshu-medical-mllm/Lingshu-32B with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="lingshu-medical-mllm/Lingshu-32B")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("lingshu-medical-mllm/Lingshu-32B")
model = AutoModelForMultimodalLM.from_pretrained("lingshu-medical-mllm/Lingshu-32B", device_map="auto")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use lingshu-medical-mllm/Lingshu-32B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "lingshu-medical-mllm/Lingshu-32B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "lingshu-medical-mllm/Lingshu-32B",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/lingshu-medical-mllm/Lingshu-32B

SGLang

How to use lingshu-medical-mllm/Lingshu-32B with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "lingshu-medical-mllm/Lingshu-32B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "lingshu-medical-mllm/Lingshu-32B",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "lingshu-medical-mllm/Lingshu-32B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "lingshu-medical-mllm/Lingshu-32B",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use lingshu-medical-mllm/Lingshu-32B with Docker Model Runner:
```
docker model run hf.co/lingshu-medical-mllm/Lingshu-32B
```

Medical report scores

by deathknight0 - opened Jun 22, 2025

Discussion

deathknight0

Jun 22, 2025

•

edited Jun 22, 2025

Thanks for the models. Your smaller model seem to outperform your larger model on most medical report benchmarks. Can you share any insights why this might be the case..?

xww033

Lingshu: MLLMs for Unified Multimodal Medical Understanding and Reasoning org Jun 23, 2025

Thanks for the models. Your smaller model seem to outperform your larger model on most medical report benchmarks. Can you share any insights why this might be the case..?

This is also very surprising to us. Since part of the training split is included in the data and the report generation data is substantially different from the majority VQA data, we hypothesize that the smaller model better fits the training set patterns of report generation. Another reason is pinpointed in our paper that current automatic evaluation metrics cannot fully reflect the model's capability of medical report generation.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment