Instructions to use roufaen/HieroSA with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use roufaen/HieroSA with Transformers:

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

pipe = pipeline("image-text-to-text", model="roufaen/HieroSA")
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, AutoModelForImageTextToText

processor = AutoProcessor.from_pretrained("roufaen/HieroSA")
model = AutoModelForImageTextToText.from_pretrained("roufaen/HieroSA")
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

vLLM

How to use roufaen/HieroSA with vLLM:

Install from pip and serve model

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

SGLang

How to use roufaen/HieroSA 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 "roufaen/HieroSA" \
    --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": "roufaen/HieroSA",
		"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 "roufaen/HieroSA" \
        --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": "roufaen/HieroSA",
		"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 roufaen/HieroSA with Docker Model Runner:
```
docker model run hf.co/roufaen/HieroSA
```

HieroSA (Chinese)

Paper | GitHub

We propose HieroSA (Hieroglyph Stroke Analyzer) 🏺, a framework for capturing stroke-level structural representations of hieroglyphic and logographic scripts. It automatically converts characters into normalized stroke-segment representations ✍️, without relying on handcrafted rules or script-specific priors.

HieroSA supports both modern logographic scripts and ancient hieroglyphs 🌍, enabling cross-lingual structural generalization. Experimental results demonstrate that it effectively captures character-level structure and semantics 🧩, providing a solid foundation for downstream analysis and understanding of hieroglyphic writing systems.

More Details

Please refer to our GitHub Repository for more details about this model, including environment setup and inference scripts.

Citation

If you find our work helpful for your research, please consider citing our work.

@article{luo2026hierosa,
    title={Enabling Stroke-Level Structural Analysis of Hieroglyphic Scripts without Language-Specific Priors}, 
    author={Fuwen Luo and Zihao Wan and Ziyue Wang and Yaluo Liu and Pau Tong Lin Xu and Xuanjia Qiao and Xiaolong Wang and Peng Li and Yang Liu},
    journal={arXiv preprint arXiv:2601.05508},
    year={2026}
}

Downloads last month: 6

Safetensors

Model size

5B params

Tensor type

BF16

Model tree for roufaen/HieroSA

Base model

Qwen/Qwen3-VL-4B-Instruct

Finetuned

(263)

this model

Paper for roufaen/HieroSA

Enabling Stroke-Level Structural Analysis of Hieroglyphic Scripts without Language-Specific Priors

Paper • 2601.05508 • Published Jan 9