使用Anaconda Promp命令在pytorch环境里运行python文件跑深度学习，实战模型LeNet使用The CIFAR-10 数据集进行分类10类运到问题及解决，且对图片进行预测

这里运行也相当于在服务器里运行深度学习，服务器中也是切换当前文件夹下运行。

1、文件位置

pycharm写的模型、训练、预测的文件名（model.py、train.py、predict.py），文件也是pytorh代码编写。
文件位置：E:\桌面\pytorch项目练习\Test1_official_demo

2、已装Anaconda，且已装pytorch环境

前提在Anaconda已装pytorch环境，打开anaconda Prompt(此处最好右键选择‘以管理员身份运行’，避免后期安装遇到权限相关警告或错误)。

(base) C:\Users\Administrator>activate pytorch38

3.进入磁盘

进入目标python文件所在磁盘：如E盘

(pytorch38) C:\Users\Administrator>E:

4.进入磁盘目标文件夹：桌面\pytorch项目练习\Test1_official_demo

(pytorch38) E:\>cd 桌面\pytorch项目练习\Test1_official_demo

5.执行目标python文件

(pytorch38) E:\桌面\pytorch项目练习\Test1_official_demo>python train.py

6.遇到问题

运行之后出现

(pytorch38) E:\桌面\pytorch项目练习\Test1_official_demo>python train.py
Files already downloaded and verified
Files already downloaded and verified
Traceback (most recent call last):
  File "train.py", line 71, in <module>
    main()
  File "train.py", line 28, in main
    val_image, val_label = val_data_iter.next()
AttributeError: '_SingleProcessDataLoaderIter' object has no attribute 'next'

7.解决

Python2 代码中迭代器的使用，引用时用*.next()的方式在Python3中不再有效，而是会报错。找到代码28行

原

val_data_iter = iter(val_loader)
    val_image, val_label = val_data_iter.next()

改后

val_data_iter = iter(val_loader)
    val_image, val_label = val_data_iter.__next__()

改后重新运行python train.py

(pytorch38) E:\桌面\pytorch项目练习\Test1_official_demo>python train.py
Files already downloaded and verified
Files already downloaded and verified
[1,   500] train_loss: 1.766  test_accuracy: 0.425
[1,  1000] train_loss: 1.442  test_accuracy: 0.522
[2,   500] train_loss: 1.235  test_accuracy: 0.578
[2,  1000] train_loss: 1.151  test_accuracy: 0.599
[3,   500] train_loss: 1.035  test_accuracy: 0.597
[3,  1000] train_loss: 1.030  test_accuracy: 0.645
[4,   500] train_loss: 0.917  test_accuracy: 0.644
[4,  1000] train_loss: 0.914  test_accuracy: 0.653
[5,   500] train_loss: 0.829  test_accuracy: 0.654
[5,  1000] train_loss: 0.842  test_accuracy: 0.659
Finished Training

8.查看

当前文件多了数据文件夹data和模型权重Lenet.pth文件

9.用一张图片测试

(pytorch38) E:\桌面\pytorch项目练习\Test1_official_demo>python predict.py
plane