深度学习碰见的报错信息汇总

1. AttributeError: module 'torchtext.data' has no attribute 'Field'

原文代码如下

import math
import time
import numpy as np
import torch
import torch.nn.functional as F 
from torchtext import data

# 两个Field对象定义字段的处理方法（文本字段、标签字段）
TEXT = data.Field(tokenize=lambda x: x.split(), lower=True)
LABEL = data.LabelField(dtype=torch.float)

更改成如下引用即可

import math
import time
import numpy as np
import torch
import torchtext
import torch.nn.functional as F 
from torchtext.legacy import data
from torchtext.legacy.data import Field, TabularDataset, Iterator, BucketIterator

2. ​​​​​​​TypeError: can't convert cuda:0 device type tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first.

问题描述
如果想把CUDA tensor格式的数据改成numpy时，需要先将其转换成cpu float-tensor随后再转到numpy格式。 numpy不能读取CUDA tensor 需要将它转化为 CPU tensor

解决办法：

self.cpu().numpy()

torch.tensor(self, device='cpu')

3. 在导入torch的时候报错

报错内容：TqdmWarning: IProgress not found. Please update jupyter and ipywidgets.

解决办法：

1. 卸载jupyter：pip uninstall jupyter
2. 卸载ipywidgets：pip uninstall ipywidgets
3. 安装jupyter：pip install jupyter
4. 安装ipywidgets：pip install ipywidgets
5. 关联：jupyter nbextension enable --py widgetsnbextension