Instructions to use adidukrembzuai/MedChain-R1-4B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use adidukrembzuai/MedChain-R1-4B with Transformers:

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

pipe = pipeline("image-text-to-text", model="adidukrembzuai/MedChain-R1-4B")
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)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("adidukrembzuai/MedChain-R1-4B")
model = AutoModelForMultimodalLM.from_pretrained("adidukrembzuai/MedChain-R1-4B", 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=256)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use adidukrembzuai/MedChain-R1-4B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "adidukrembzuai/MedChain-R1-4B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "adidukrembzuai/MedChain-R1-4B",
		"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/adidukrembzuai/MedChain-R1-4B

SGLang

How to use adidukrembzuai/MedChain-R1-4B 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 "adidukrembzuai/MedChain-R1-4B" \
    --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": "adidukrembzuai/MedChain-R1-4B",
		"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 "adidukrembzuai/MedChain-R1-4B" \
        --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": "adidukrembzuai/MedChain-R1-4B",
		"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 adidukrembzuai/MedChain-R1-4B with Docker Model Runner:
```
docker model run hf.co/adidukrembzuai/MedChain-R1-4B
```

adidukrembzuai commited on 3 days ago

Commit

22b4cf4

verified ·

1 Parent(s): 0ff4483

Overview: Figure 1 from arXiv:2611.00041v1, with caption

Browse files

Files changed (2) hide show

README.md +6 -0
overview.png +2 -2

README.md CHANGED Viewed

@@ -29,6 +29,12 @@ tags:
   <img src="overview.png" alt="MedChain-R1 overview" width="900"/>
 </p>
 **MedChain-R1-4B** is a 4B vision-language model for **sequential medical image grounding**: given a short sequence of
 medical images and a clinical question, it first writes an image-grounded rationale in `<think>…</think>` and then
 returns one bounding box `<answer>(x1,y1),(x2,y2)</answer>` in pixel coordinates. It is built on

   <img src="overview.png" alt="MedChain-R1 overview" width="900"/>
 </p>
+*Figure 1 of the paper: **Sequential medical grounding with reasoning supervision.** (a) The eight MedSG-Bench tasks,
+grouped into image difference grounding (IDG) and image consistency grounding (ICG). (b) A registered-difference example:
+zero-shot Qwen3.5 and MedSeq-Grounder return imprecise boxes, whereas MedChain-R1 first identifies the new nodule and then
+localizes it. (c) MedSG-Bench mIoU and Acc@0.5 of MedSeq-Grounder (7B) and MedChain-R1 at 4B and 9B (dCoT-SFT and
+dCoT-SFT+GRPO) and 27B (dCoT-SFT).*
 **MedChain-R1-4B** is a 4B vision-language model for **sequential medical image grounding**: given a short sequence of
 medical images and a clinical question, it first writes an image-grounded rationale in `<think>…</think>` and then
 returns one bounding box `<answer>(x1,y1),(x2,y2)</answer>` in pixel coordinates. It is built on

overview.png CHANGED Viewed

Git LFS Details

SHA256: dde135fc315ad4513badbcca32e75b16f65e65c5ed03f6ecf6e4acc7dc3668a0
Pointer size: 131 Bytes
Size of remote file: 798 kB

Git LFS Details

SHA256: 1c9218f5b1c78d673c6963e657de73b1cf59d8ecd888d0614bc76158da4dde0a
Pointer size: 131 Bytes
Size of remote file: 160 kB