cloudyskyy / Guwen-UNILM
本仓库是基于bert4keras实现的古文-现代文翻译模型。具体使用了基于掩码自注意力机制的UNILM(Li al., 2019)预训练模型作为翻译系统的backbone。我们首先使用了普通的中文（现代文）BERT、Roberta权重作为UNILM的初始权重以训练UNILM模型（具体在文中分别为B-UNILM以及R-UNILM）。为了更好的使UNILM模型适应古文的特性，我们尝试使用了在古文预训练模型Guwen-BERT，作为UNILM的初始权重，并且获得了最优的效果。
☆42Updated 2 years ago
Related projects ⓘ
Alternatives and complementary repositories for Guwen-UNILM
- SikuBERT：四库全书的预训练语言模型（四库BERT） Pre-training Model of Siku Quanshu☆115Updated last year
- A Benchmark for Classical Chinese Based on a Crowdsourcing System.☆50Updated 3 years ago
- 古文语言理解测评基准 Classical Chinese Language Understanding Evaluation Benchmark: datasets, baselines, pre-trained models, corpus and leaderboard☆46Updated last year
- 一个面向繁体中文古籍分词的python工具包☆31Updated 2 years ago
- ☆31Updated last year
- Yet Another Chinese Learner Corpus☆72Updated 2 years ago
- 基于SpanBert的中文指代消解，pytorch实现☆95Updated last year
- 历届中文句法错误诊断技术评测数据集☆35Updated 2 years ago
- CCL 2022 汉语学习者文本纠错评测☆135Updated last year
- SIGHAN中文纠错数据集及转换后格式☆63Updated 4 years ago
- GuwenModels: 古文自然语言处理模型合集, 收录互联网上的古文相关模型及资源. A collection of Classical Chinese natural language processing models, including Classical Ch…☆159Updated 11 months ago
- 基于pytorch的中文拼写纠错，使用的模型是Bert以及SoftMaskedBert☆31Updated 3 years ago
- CCL2022汉语学习者文本纠错评测任务赛道二——CGED-8第一名解决方案☆52Updated last year
- <u><a href="https://circse.github.io/LT4HALA/" style="color: white">Workshop on Language Technologies for Historical and Ancient Language…☆32Updated 5 months ago
- ChineseBert用于中文拼写纠错☆39Updated last year
- 中文机器 阅读理解数据集☆100Updated 3 years ago
- [Unofficial] Predict code for AAAI 2022 paper: Unified Named Entity Recognition as Word-Word Relation Classification☆50Updated 2 years ago
- 基于pytorch的GlobalPointer进行中文命名实体识别。☆36Updated last year
- ☆21Updated last year
- 基于prompt的中文文本分类。☆55Updated last year
- Code and data of the paper "MCTS: A Multi-Reference Chinese Text Simplification Dataset".☆28Updated 5 months ago
- ☆16Updated 2 years ago
- 基于BERT和MRC框架实现的嵌套命名实体识别☆18Updated 2 years ago
- 基于T5模型的中文文本纠错☆26Updated 2 weeks ago
- Dynamic Connected Networks for Chinese Spelling Check☆49Updated 7 months ago
- 大规模中文自然语言处理语料 Large Scale Chinese Corpus for NLP☆24Updated 5 years ago
- 百度2021年语言与智能技术竞赛多形态信息抽取赛道关系抽取部分torch版baseline☆50Updated 3 years ago
- Code & data for our EMNLP2022 paper "SynGEC: Syntax-Enhanced Grammatical Error Correction with a Tailored GEC-Oriented Parser"☆80Updated 8 months ago
- 非官方的MDCSpell论文的实现☆17Updated 2 years ago
- 基于seq2edit (Gector) 的中文文本纠错。☆26Updated 2 years ago