We first pre-trained Transformer-XL on the Wudo corpus (180G version), and then fine-tuned it on a denoised dataset (developed by us). The denoise task is to reconstruct a fluent and clean text from a noisy input which includes random insertion/swap/deletion/replacement/sentence reordering.

使用 Usage

加载模型 Loading Models

git clone https://github.com/IDEA-CCNL/Fengshenbang-LM.git

python

from fengshen.models.transfo_xl_denoise.tokenization_transfo_xl_denoise import TransfoXLDenoiseTokenizer
from fengshen.models.transfo_xl_denoise.modeling_transfo_xl_denoise import TransfoXLDenoiseModel

tokenizer = TransfoXLDenoiseTokenizer.from_pretrained('IDEA-CCNL/Randeng-Transformer-1.1B-Denoise')
model = TransfoXLDenoiseModel.from_pretrained('IDEA-CCNL/Randeng-Transformer-1.1B-Denoise')

使用示例 Usage Examples

python

from fengshen.models.transfo_xl_denoise.generate import denoise_generate
input_text = "凡是有成就的人, 都很严肃地对待生命自己的"
res = denoise_generate(model, tokenizer,  input_text)
print(res)
# "有成就的人都很严肃地对待自己的生命。"

引用 Citation

如果您在您的工作中使用了我们的模型，可以引用我们的论文：

If you are using the resource for your work, please cite the our paper:

text

@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}
}

也可以引用我们的网站:

You can also cite our website:

text

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