Instructions to use OpenDFM/CharTool-3B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OpenDFM/CharTool-3B with Transformers:

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

pipe = pipeline("image-text-to-text", model="OpenDFM/CharTool-3B")
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("OpenDFM/CharTool-3B")
model = AutoModelForMultimodalLM.from_pretrained("OpenDFM/CharTool-3B", 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]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use OpenDFM/CharTool-3B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "OpenDFM/CharTool-3B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "OpenDFM/CharTool-3B",
		"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/OpenDFM/CharTool-3B

SGLang

How to use OpenDFM/CharTool-3B 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 "OpenDFM/CharTool-3B" \
    --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": "OpenDFM/CharTool-3B",
		"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 "OpenDFM/CharTool-3B" \
        --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": "OpenDFM/CharTool-3B",
		"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 OpenDFM/CharTool-3B with Docker Model Runner:
```
docker model run hf.co/OpenDFM/CharTool-3B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

CharTool-3B

Paper | Code | CharTool-7B

CharTool-3B is a tool-integrated multimodal agent for fine-grained chart perception and accurate numerical reasoning. It can use image cropping for localized visual perception and Python code execution for numerical computation.

📋 Model Details

Base model: Qwen2.5-VL-3B-Instruct
Training: Cold-start supervised fine-tuning followed by agentic reinforcement learning on DuoChart
Parameters: 3B

🚀 Usage

CharTool relies on a tool-integrated inference loop and a code sandbox. Please follow the evaluation instructions in the CharTool repository.

Use this Hugging Face repository as the model path:

python src/generate.py \
    --model_name chartool \
    --split val \
    --mode reasoning \
    --model_path OpenDFM/CharTool-3B

📝 Citation

@article{zhang2026chartool,
  title   = {CharTool: Tool-Integrated Visual Reasoning for Chart Understanding},
  author  = {Zhang, Situo and Zhang, Yifan and Zhu, Zichen and Ma, Da and Pan, Lei and Zhang, Danyang and Zhao, Zihan and Chen, Lu and Yu, Kai},
  journal = {arXiv preprint arXiv:2604.02794},
  year    = {2026}
}