Edit model card
YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/model-cards#model-card-metadata)

ChineseBERT-large

This repository contains code, model, dataset for ChineseBERT at ACL2021.

paper:
ChineseBERT: Chinese Pretraining Enhanced by Glyph and Pinyin Information
Zijun Sun, Xiaoya Li, Xiaofei Sun, Yuxian Meng, Xiang Ao, Qing He, Fei Wu and Jiwei Li

code:
ChineseBERT github link

Model description

We propose ChineseBERT, which incorporates both the glyph and pinyin information of Chinese characters into language model pretraining.

First, for each Chinese character, we get three kind of embedding.

  • Char Embedding: the same as origin BERT token embedding.
  • Glyph Embedding: capture visual features based on different fonts of a Chinese character.
  • Pinyin Embedding: capture phonetic feature from the pinyin sequence ot a Chinese Character.

Then, char embedding, glyph embedding and pinyin embedding are first concatenated, and mapped to a D-dimensional embedding through a fully connected layer to form the fusion embedding.
Finally, the fusion embedding is added with the position embedding, which is fed as input to the BERT model.
The following image shows an overview architecture of ChineseBERT model.

MODEL

ChineseBERT leverages the glyph and pinyin information of Chinese characters to enhance the model's ability of capturing context semantics from surface character forms and disambiguating polyphonic characters in Chinese.

Downloads last month
27
Unable to determine this model’s pipeline type. Check the docs .