Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/foruck/sentiment-analysis-with-base-bert
Project for SJTU CS438
https://github.com/foruck/sentiment-analysis-with-base-bert
Last synced: about 1 month ago
JSON representation
Project for SJTU CS438
- Host: GitHub
- URL: https://github.com/foruck/sentiment-analysis-with-base-bert
- Owner: Foruck
- License: gpl-3.0
- Created: 2018-11-21T10:22:48.000Z (about 6 years ago)
- Default Branch: master
- Last Pushed: 2019-10-03T07:27:07.000Z (over 5 years ago)
- Last Synced: 2023-09-29T16:07:03.795Z (over 1 year ago)
- Language: Jupyter Notebook
- Homepage:
- Size: 16.6 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
- Code of conduct: CODE_OF_CONDUCT.md
Awesome Lists containing this project
README
# Sentiment Analysis with BERT
Project for SJTU CS438
## 方法
- 中文:BERT + TextCNN + Bi-LSTM + Self-attention
- 英文:BERT + Bi-LSTM + Self-attention
## 实验
中文34639条,英文13385条
### 预处理:preprocess_bert.py
BERT分词
### 训练:bert_analysis.ipynb
用80%作为train集合,20%作为valid集合。
BertAdam,分类器学习率1e-3,微调学习率5e-5, batch大小24,进行30个epoch,结果取最优。
valid:中文89.68% 英文89.84%
### 测试:evaluate.ipynb
中英文各5000条
中文82+% 英文85+%
## To do
- 用机器翻译方法扩充数据集。
- 尝试Capsule、xgboost等其他分类器结构。
- BERT分层加权
By 刘欣鹏