Wobert Chinese Base

🚀 WoBERT

This project provides the TF and PyTorch versions of WoBERT, which is a word-based Chinese BERT model.

🚀 Quick Start

📦 Installation (Mainly for Installing WoBertTokenizer)

⚠️ Important Note

The version of transformers needs to be >= 4.7.0.

The implementation of WoBertTokenizer is the same as that of RoFormerTokenizer, so you can use RoFormerTokenizer directly.

💻 Usage Examples

Basic Usage

import torch
from transformers import BertForMaskedLM as WoBertForMaskedLM
from transformers import RoFormerTokenizer as WoBertTokenizer

pretrained_model_or_path_list = [
    "junnyu/wobert_chinese_plus_base", "junnyu/wobert_chinese_base"
]
for path in pretrained_model_or_path_list:
    text = "今天[MASK]很好，我[MASK]去公园玩。"
    tokenizer = WoBertTokenizer.from_pretrained(path)
    model = WoBertForMaskedLM.from_pretrained(path)
    inputs = tokenizer(text, return_tensors="pt")
    with torch.no_grad():
        outputs = model(**inputs).logits[0]
    outputs_sentence = ""
    for i, id in enumerate(tokenizer.encode(text)):
        if id == tokenizer.mask_token_id:
            tokens = tokenizer.convert_ids_to_tokens(outputs[i].topk(k=5)[1])
            outputs_sentence += "[" + "||".join(tokens) + "]"
        else:
            outputs_sentence += "".join(
                tokenizer.convert_ids_to_tokens([id],
                                                skip_special_tokens=True))
    print(outputs_sentence)
# RoFormer    今天[天气||天||心情||阳光||空气]很好，我[想||要||打算||准备||喜欢]去公园玩。
# PLUS WoBERT 今天[天气||阳光||天||心情||空气]很好，我[想||要||打算||准备||就]去公园玩。
# WoBERT      今天[天气||阳光||天||心情||空气]很好，我[想||要||就||准备||也]去公园玩。

📚 Documentation

TF Version

https://github.com/ZhuiyiTechnology/WoBERT

PyTorch Version

https://github.com/JunnYu/WoBERT_pytorch

📄 License

No license information provided in the original document.

📖 Citation

Bibtex:

@techreport{zhuiyiwobert,
  title={WoBERT: Word-based Chinese BERT model - ZhuiyiAI},
  author={Jianlin Su},
  year={2020},
  url="https://github.com/ZhuiyiTechnology/WoBERT",
}