Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

Appearance settings

BasicCoder/Weibo-Sentiment-Analysis

Open more actions menu

Repository files navigation

Weibo Sentiment Analysis

  1. 基于开源代码中改写的
  2. 大多数情感分析模型只是开源代码,没有提供对应预训练模型。针对微博情感分析同样也缺乏相应的预训练模型,本模型提供情感分析训练代码,以及模型推断
  3. 主要针对的数据集
    数据集 类型 链接
    weibo_senti_100k 2种感情,正向评论和负向评论 地址
    simplifyweibo_4_moods 4 种情感,喜悦、愤怒、厌恶、低落 地址

Environmental preparation

This repo was tested on 3.6 and PyTorch 0.4.1/1.0.0

pip install pytorch-pretrained-bert

Model Train

采用 BERT-base, ChineseGoogle 模型在数据集上进行fine-tune

准备数据集

划分 Train、Valid、Test 数据集

cd Weibo-Sentiment-Analysis
python split_weibo_senti_100k.py

模型训练

sh train.sh
模型 数据集 F1
Bert-base weibo_senti_100k 0.9840746441110737
simplifyweibo_4_moods 0.4969952552559987
Bert-wwm-ext --
XLNet-base --
XLNet-mid --
XLNet-large --
Roberta-mid --

Model Test

预训练模型下载:

模型 数据集 下载地址
Bert-base weibo_senti_100k 地址
simplifyweibo_4_moods 地址

将下载下来的模型文件保存在./checkpoints 目录下

sh test.sh

About

Sentiment Analysis for weibo-sent-100k

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Morty Proxy This is a proxified and sanitized view of the page, visit original site.