(一) PyTorch实战CIFAR10数据_读取与展示

（1）导包

import torch
import torchvision
import torchvision.transforms as transforms
from torch.utils.data import DataLoader
import matplotlib.pyplot as plt
import numpy as np

 The torchvision package consists of popular datasets, model architectures, and common image transformations for computer vision.

（2）读取数据

transform = transforms.Compose(
    [transforms.ToTensor(),
     transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))] # 彩色图像三通道 均值 std 标准差
)

trainset = torchvision.datasets.CIFAR10(root='./data', train=True, download=True, transform=transform)
trainloader = DataLoader(trainset, batch_size=4, shuffle=True, num_workers=2)
testset = torchvision.datasets.CIFAR10(root='./data', train=False, download=True, transform=transform)
testloader = DataLoader(testset, batch_size=4, shuffle=True, num_workers=2)

（3）展示某一个批次的图像和类别

def imshow(img):
    img = img / 2 + 0.5
    npim = img.numpy() # 张量转化成array
    plt.imshow(np.transpose(npim, (1, 2, 0)))
    plt.show()

if __name__ == '__main__':
    # 随机获取部分训练数据
    dataiter = iter(trainloader)
    images, labels = dataiter.next();

    # 显示图片
    imshow(torchvision.utils.make_grid(images)) # make_grid 返回张量类型
    classes = ('plane', 'car', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'trunk')
    print(' '.join('%5s' % classes[labels[j]] for j in range(4)))

 （4）可能的问题

1）报错原因

上述错误是因为在Windows环境下在读取数据时用DataLoader包装数据时，用多线程读取数据会出bug 。

2）解决方法

a) 将线程num_worker设置为0 

trainset = torchvision.datasets.CIFAR10(root='./data', train=True, download=False, transform=transform) # 从root中读取数据，设置train和transform
trainloader = DataLoader(trainset, batch_size=4, shuffle=True, num_workers=0) # 包装数据 num_workers 两个线程读取数据

 b) 添加 if __name__ == '__main': # 程序入口

if __name__ == '__main__':
    # 随机获取部分训练数据
    dataiter = iter(trainloader)
    images, labels = dataiter.next();

    # 显示图片
    imshow(torchvision.utils.make_grid(images)) # make_grid 返回张量类型
    print(' '.join('%5s' % classes[labels[j]] for j in range(4)))

 （5）参考资料

torchvisionhttps://pytorch.org/vision/stable/index.html

RuntimeError：An attempt has been made to start a new process before the current process has finished its bootstrapping phase.https://blog.csdn.net/u013700358/article/details/82753019      Python 中的 if __name__ == '__main__' 该如何理解https://blog.konghy.cn/2017/04/24/python-entry-program/