Bert-Chinese-Text-Classification-Pytorch issues

关于自己数据集的问题

1

1.在读完代码后想换成自己的数据集，请问用什么工具可以做成这个数据集的标准模式。 2.测试代码和验证代码我看都写好了，请问一下怎么跑测试代码和验证代码呢。萌新出来驾到，还希望大佬们解答一下。谢谢你们了。

cc846049716

运行错误 AttributeError: 'NoneType' object has no attribute 'tokenize'

3

AttributeError: 'NoneType' object has no attribute 'tokenize' 求救

Upsetin

ModuleNotFoundError: No module named 'fused_layer_norm_cuda'

2

![image](https://user-images.githubusercontent.com/24981824/139167922-1caf9ec2-9363-4d97-8554-0c9aa8bb2975.png) 这是什么问题呀

scutfrank

数据格式是这种短句 ![image](https://user-images.githubusercontent.com/62530854/124207825-e99cd680-db18-11eb-9c57-9bb5b411dc92.png) 下面这个是属于什么问题呢，改怎么改呢？用的bert模型，没有修改参数 ![image](https://user-images.githubusercontent.com/62530854/124207839-eefa2100-db18-11eb-85fa-d17467ecfd87.png)

cfy42584125

size mismatch for bert.embeddings.LayerNorm.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([768]).

![image](https://user-images.githubusercontent.com/59283713/135701802-537108cc-c80b-4abf-9c37-ec9ad53904bd.png) 很奇怪，不知道怎么解决更改self.hidden_size = 768也没有用

dongandi

ZMM6128

init_network

想问一下网络没有进行初始化吗

Aronbryant

Bert-Chinese-Text-Classification-Pytorch
Bert-Chinese-Text-Classification-Pytorch copied to clipboard

Metadata

关于自己数据集的问题

运行错误 AttributeError: 'NoneType' object has no attribute 'tokenize'

ModuleNotFoundError: No module named 'fused_layer_norm_cuda'

Update README.md

改用自己的数据集，训练效果问题

size mismatch for bert.embeddings.LayerNorm.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([768]).

同一个字不加##和加##区别是什么？

训练迭代30次，就train_acc 就100%了

您好！请问为什么会出现类的个数跟target_name不匹配的问题呢请问有什么解决办法

init_network

← Metadata

Owner

Metadata

Bert-Chinese-Text-Classification-Pytorch Bert-Chinese-Text-Classification-Pytorch copied to clipboard

Metadata

← Metadata

Owner

Metadata

Bert-Chinese-Text-Classification-Pytorch
Bert-Chinese-Text-Classification-Pytorch copied to clipboard