论文笔记：Enhancing Pre-trained Chinese Character Representation with Word-aligned Attention
1. 概述 目前，很多NLP算法大多采用主流的预训练模型+下游任务微调这样的算法架构。预训练模型种类繁多，如下图 用的最多的莫过于大名鼎鼎的BERT预训练模型，同样是基于Pre-training和Fine-tuning模式架构的 不管啥模型，第一件事都是 tokenizer。对于 BERT 来说，英文的 token 是 word-piece，中文的是字（这也对后面的实验造成了很大的麻烦，因为要对齐
相关文章
- 1. Character-Level Question Answering with Attention 论文笔记
- 2. 【论文笔记】Knowledgeable Reader_ Enhancing Cloze-Style Reading Comprehension with External Commonsense Kg
- 3. 论文笔记：Image Captioning with Semantic Attention
- 4. 【论文笔记】Enhancing Pre-Trained Language Representations with Rich Knowledge for MRC
- 5. 论文：Character-Level Language Modeling with Deeper Self-Attention
- 6. 论文笔记：Representation Learning with Contrastive Predictive Coding
- 7. Character-Level Question Answering with Attention读书笔记
- 8. 论文《Chinese Poetry Generation with Recurrent Neural Network》阅读笔记
- 9. 【论文笔记】QANET：Combining Local Convolution With Global Self-attention for Reading Comprehension
- 10. 【论文笔记】Reasoning about Entailment with Neural Attention
- 更多相关文章...
- • Swift 字符(Character) - Swift 教程
- • XSL-FO character 对象 - XSL-FO 教程
- • Tomcat学习笔记（史上最全tomcat学习笔记）
- • Scala 中文乱码解决
相关标签/搜索
每日一句
-
每一个你不满意的现在，都有一个你没有努力的曾经。
欢迎关注本站公众号,获取更多信息
相关文章
- 1. Character-Level Question Answering with Attention 论文笔记
- 2. 【论文笔记】Knowledgeable Reader_ Enhancing Cloze-Style Reading Comprehension with External Commonsense Kg
- 3. 论文笔记：Image Captioning with Semantic Attention
- 4. 【论文笔记】Enhancing Pre-Trained Language Representations with Rich Knowledge for MRC
- 5. 论文：Character-Level Language Modeling with Deeper Self-Attention
- 6. 论文笔记：Representation Learning with Contrastive Predictive Coding
- 7. Character-Level Question Answering with Attention读书笔记
- 8. 论文《Chinese Poetry Generation with Recurrent Neural Network》阅读笔记
- 9. 【论文笔记】QANET：Combining Local Convolution With Global Self-attention for Reading Comprehension
- 10. 【论文笔记】Reasoning about Entailment with Neural Attention