README.md

---
license: apache-2.0
---

# mirror-chinese-mrcqa-alpha

🎉 Congratulations, you've found something fantastic!

This is a ckpt of [🪞Mirror](https://github.com/Spico197/Mirror) for Chinese language.

This model supports a lot of schema-guided Chinese Information Extraction tasks, e.g. NER, RE, Keyword extraction, MRC, Classification.

The model is developed in a very early stage (before the final Mirror) and it is based on MRC.

Please contact [Tong Zhu](https://spico197.github.io/) for more information about the model usage.

## 🚀 QuickStart

- downloading

```bash
git clone https://github.com/Spico197/Mirror
# Mirror is the working dir
cd Mirror
mkdir outputs
cd outputs
git lfs install
git clone https://huggingface.co/Spico/mirror-chinese-mrcqa-alpha
```

- usage

```bash
# 1. go to Mirror (the working dir)
# 2. change `src/app/demo1_deprecated.py`: `MrcQaPipeline("outputs/RobertaBase_data20230314v2")` to `MrcQaPipeline("outputs/mirror-chinese-mrcqa-alpha")`
python -m src.app.demo1_deprecated
```

## 📖 Citation

```bibtex
@misc{zhu_mirror_2023,
  shorttitle = {Mirror},
  title = {Mirror: A Universal Framework for Various Information Extraction Tasks},
  author = {Zhu, Tong and Ren, Junfei and Yu, Zijian and Wu, Mengsong and Zhang, Guoliang and Qu, Xiaoye and Chen, Wenliang and Wang, Zhefeng and Huai, Baoxing and Zhang, Min},
  url = {http://arxiv.org/abs/2311.05419},
  doi = {10.48550/arXiv.2311.05419},
  urldate = {2023-11-10},
  publisher = {arXiv},
  month = nov,
  year = {2023},
  note = {arXiv:2311.05419 [cs]},
  keywords = {Computer Science - Artificial Intelligence, Computer Science - Computation and Language},
}
```