Bert-Chinese-Text-Classification-Pytorch icon indicating copy to clipboard operation
Bert-Chinese-Text-Classification-Pytorch copied to clipboard

使用Bert,ERNIE,进行中文文本分类

Results 145 Bert-Chinese-Text-Classification-Pytorch issues
Sort by recently updated
recently updated
newest added

如果最后的结果只有总体的准确率,F1分数等等,那我训练他进行具体文本分类的意义在哪里?

通过bert训练后字向量维度是768吗?

The model is not available: http://image.nghuyong.top/ERNIE.zip {"error":"Document not found"} Also, google drive is not accessible in China. Could you please provide another link?

第一轮训练完52分钟,train acc :46% ,val acc:50% 第二轮训练完居然也只有第一轮这个准确率 完全按照作者的步骤来的,不知道哪里除了问题,麻烦作者或者其他大佬有没有解决方案

报错信息: Traceback (most recent call last): File "/baykal/public/zhangmeiyu/bert/run.py", line 39, in model = x.Model(config).to(config.device) File "/baykal/public/zhangmeiyu/bert/models/bert.py", line 39, in __init__ self.bert = BertModel.from_pretrained(config.bert_path) File "/baykal/public/zhangmeiyu/bert/pytorch_pretrained/modeling.py", line 604, in from_pretrained state_dict...

請問要改繁體訓練資料 vocab.txt pytorch_model.bin bert_config.json 要如何取得 或變更 謝謝

代码中masked language modeling labels中-1标记的是被masked的token,loss计算忽略被mask的token,但是BERT论文中写的是”the final hidden vectors corresponding to the mask tokens are fed into an output softmax over the vocabulary“ 只计算masked token处的loss

项目开发者未曾提到包还需要安装boto3 >sklearn的conda包名为scikit-learn 项目可运行的环境 [requirement.txt](https://github.com/649453932/Bert-Chinese-Text-Classification-Pytorch/files/10793979/requirement.txt) 创建虚拟环境时,使用该命令 ```CONSOLE conda env create -f C:\XXX\environment.txt ``` 即可直接上手