Chinesebert Large

🚀 ChineseBert_pytorch

This project mainly customizes the ChineseBertTokenizerFast code in the tokenization_chinesebert_fast.py file, enabling it to be called from huggingface.co.

🚀 Quick Start

This project mainly customizes the ChineseBertTokenizerFast code in the tokenization_chinesebert_fast.py file, so it can be called from huggingface.co.

pretrained_tokenizer_name = "junnyu/ChineseBERT-large"
tokenizer = ChineseBertTokenizerFast.from_pretrained(pretrained_tokenizer_name)

📚 Documentation

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

📦 Installation

pip install chinesebert
or
pip install git+https://github.com/JunnYu/ChineseBert_pytorch.git

💻 Usage Examples

Basic Usage

import torch
from chinesebert import ChineseBertForMaskedLM, ChineseBertTokenizerFast, ChineseBertConfig
pretrained_model_name = "junnyu/ChineseBERT-large"
tokenizer = ChineseBertTokenizerFast.from_pretrained(pretrained_model_name )
chinese_bert = ChineseBertForMaskedLM.from_pretrained(pretrained_model_name)

text = "北京是[MASK]国的首都。"
inputs = tokenizer(text, return_tensors="pt")
print(inputs)
maskpos = 4

with torch.no_grad():
    o = chinese_bert(**inputs)
    value, index = o.logits.softmax(-1)[0, maskpos].topk(10)

pred_tokens = tokenizer.convert_ids_to_tokens(index.tolist())
pred_values = value.tolist()

outputs = []
for t, p in zip(pred_tokens, pred_values):
    outputs.append(f"{t}|{round(p,4)}")
print(outputs)

# base  ['中|0.711', '我|0.2488', '祖|0.016', '法|0.0057', '美|0.0048', '全|0.0042', '韩|0.0015', '英|0.0011', '两|0.0008', '王|0.0006']
# large ['中|0.8341', '我|0.1479', '祖|0.0157', '全|0.0007', '国|0.0005', '帝|0.0001', '该|0.0001', '法|0.0001', '一|0.0001', '咱|0.0001']

📄 Reference

https://github.com/ShannonAI/ChineseBert