Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.
You can not select more than 25 topics
Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.
houysh
045052c9d3
|
2 years ago | |
|---|---|---|
| text_classification | 2 years ago | |
| README.md | 2 years ago | |
| model_comp.png | 2 years ago | |
README.md
基于TensorLayer的NLP算法
文本分类
目前基于TensorLayer接口已实现基于RNN、LSTM、GRU、TransformerEncoder等网络结构的文本分类算法。
运行环境
- TensorLayer3
- Backend：TensorLayer运行所需的后端，如TensorFlow、Pytorch、Paddle
- TensorLayer使用手册
使用
- 模型训练
python text_classification_tl.py --run_mode train --network lstm --epochs=2 --model_path model_lstm.npz
- 模型评估
python text_classification_tl.py --run_mode eval --network lstm --model_path model_lstm.npz
实验
任务
情感分析
数据集
| 标签 | 样本数 |
|---|---|
| 0 | 12500 |
| 1 | 12500 |
随机抽取20%的数据作为测试集，剩余80%的数据为训练集
模型对比
| 模型 | 精确率 |
|---|---|
| RNN | 0.8544 |
| LSTM | 0.8534 |
| GRU | 0.8548 |
| TransformerEncoder | 0.8586 |
