Randeng Pegasus 238M Chinese

🚀 Randeng-Pegasus-238M-Chinese

A Chinese version of PAGASUS-base, specialized in text summarization tasks.

• Main Page: Fengshenbang
• Github: Fengshenbang-LM

🚀 Quick Start

Randeng-Pegasus-238M-Chinese is a Chinese version of PAGASUS-base, designed to excel in text summarization tasks.

✨ Features

• Task-Specific: Specialized for Chinese text summarization.
• Multi-Version Support: Offers both large and fine-tuned versions.

📦 Installation

The installation mainly involves setting up the necessary Python environment and downloading the model. You need to download tokenizers_pegasus.py and other Python scripts in advance. You can either get them from the Fengshenbang-LM GitHub repository or from here. It is strongly recommended to clone the Fengshenbang-LM repository:

git clone https://github.com/IDEA-CCNL/Fengshenbang-LM
cd Fengshenbang-LM/fengshen/examples/pegasus/

💻 Usage Examples

Basic Usage

from transformers import PegasusForConditionalGeneration
# Need to download tokenizers_pegasus.py and other Python script from Fengshenbang-LM github repo in advance,
# or you can download tokenizers_pegasus.py and data_utils.py in https://huggingface.co/IDEA-CCNL/Randeng_Pegasus_238M/tree/main
# Stronly recomend you git clone the Fengshenbang-LM repo:
# 1. git clone https://github.com/IDEA-CCNL/Fengshenbang-LM
# 2. cd Fengshenbang-LM/fengshen/examples/pegasus/
# and then you will see the tokenizers_pegasus.py and data_utils.py which are needed by pegasus model
from tokenizers_pegasus import PegasusTokenizer

model = PegasusForConditionalGeneration.from_pretrained("IDEA-CCNL/Randeng-Pegasus-238M-Chinese")
tokenizer = PegasusTokenizer.from_pretrained("IDEA-CCNL/Randeng-Pegasus-238M-Chinese")

text = "据微信公众号“界面”报道，4日上午10点左右，中国发改委反垄断调查小组突击查访奔驰上海办事处，调取数据材料，并对多名奔驰高管进行了约谈。截止昨日晚9点，包括北京梅赛德斯-奔驰销售服务有限公司东区总经理在内的多名管理人员仍留在上海办公室内"
inputs = tokenizer(text, max_length=512, return_tensors="pt")

# Generate Summary
summary_ids = model.generate(inputs["input_ids"])
tokenizer.batch_decode(summary_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
# model output: 截止昨日晚9点，包括北京梅赛德斯-奔驰销售服务有限公司东区总经理在内的多名管理人员仍留在上海办公室内

📚 Documentation

Model Taxonomy

Model Information

• Reference Paper: PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization
• Training Approach: To address Chinese abstractive summarization tasks, we follow the PEGASUS design. We use a version of the WuDao Corpora (180 GB) for pre-training. Considering the instability of Chinese sentence segmentation, we use both Jieba and BERT tokenizers in Randeng-PEGASUS.
• Other Versions: We also offer a large version IDEA-CCNL/Randeng-Pegasus-523M-Chinese and a fine-tuned version on Chinese summarization datasets Randeng-Pegasus-238M-Summary-Chinese.

📄 License

If you use our model in your work, please cite our paper:

@article{fengshenbang,
  author    = {Jiaxing Zhang and Ruyi Gan and Junjie Wang and Yuxiang Zhang and Lin Zhang and Ping Yang and Xinyu Gao and Ziwei Wu and Xiaoqun Dong and Junqing He and Jianheng Zhuo and Qi Yang and Yongfeng Huang and Xiayu Li and Yanghan Wu and Junyu Lu and Xinyu Zhu and Weifeng Chen and Ting Han and Kunhao Pan and Rui Wang and Hao Wang and Xiaojun Wu and Zhongshen Zeng and Chongpei Chen},
  title     = {Fengshenbang 1.0: Being the Foundation of Chinese Cognitive Intelligence},
  journal   = {CoRR},
  volume    = {abs/2209.02970},
  year      = {2022}
}

You can also cite our website:

@misc{Fengshenbang-LM,
  title={Fengshenbang-LM},
  author={IDEA-CCNL},
  year={2021},
  howpublished={\url{https://github.com/IDEA-CCNL/Fengshenbang-LM}},
}