suolyer's picture
Update README.md
e760151
metadata
language:
  - zh
license: apache-2.0
widget:
  - text: 生活的真谛是[MASK]。

longformer model (Chinese),one model of Fengshenbang-LM.

We modify the original position code of longformer to rotational position coding,and on the basis of chinese_roformer_L-12_H-768_A-12.zip, use 180G of data to continue training

Usage

There is no structure of Longformer-base in Transformers, you can run follow code to get structure of longformer from Fengshenbang-LM

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

Load Model

from fengshen import LongformerModel    
from fengshen import LongformerConfig
from transformers import BertTokenizer

tokenizer = BertTokenizer.from_pretrained("IDEA-CCNL/longformer_base")
config = LongformerConfig.from_pretrained("IDEA-CCNL/longformer_base")
model = LongformerModel.from_pretrained("IDEA-CCNL/longformer_base")

Citation

If you find the resource is useful, please cite the following website in your paper.

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