MNIST数据集转为图片形式输出（按标签分别存储，统一存储） 好用！

统一存储

将MNIST中所有的图片数据不区分索引，存储在data_train、data_test文件夹中。

废话不多说，直接代码展示：

import os
import torch
import torchvision
import torchvision.transforms as transforms
from skimage import io
import torchvision.datasets.mnist as mnist

# Device configuration
device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')

# MNIST dataset
train_dataset = torchvision.datasets.MNIST(root='data/',
                                           train=True,
                                           transform=transforms.ToTensor(),
                                           download=True)

test_dataset = torchvision.datasets.MNIST(root='data/',
                                          train=False,
                                          transform=transforms.ToTensor())

root='./data/MNIST/raw/'

train_set=(mnist.read_image_file(os.path.join(root,'train-images-idx3-ubyte')),
           mnist.read_label_file(os.path.join(root,'train-labels-idx1-ubyte'))
           )
test_set = (
    mnist.read_image_file(os.path.join(root,'t10k-images-idx3-ubyte')),
    mnist.read_label_file(os.path.join(root,'t10k-labels-idx1-ubyte'))
)

def convert_to_img(train=True):
    if(train):
        f=open('./data/train.txt','w')
        data_path='./data/data_train/'
        if(not os.path.exists(data_path)):
            os.makedirs(data_path)
        for i,(img,label) in enumerate(zip(train_set[0],train_set[1])):
            img_path=data_path+str(i)+'.jpg'
            print('train_img_path:', img_path, 'train_img_num:', i)
            io.imsave(img_path,img.numpy())
            f.write(str(label.item()) + '\n')
        f.close()
    else:
        # f = open(root + 'test.txt', 'w')
        f = open('./data/test.txt', 'w')
        data_path = './data/data_test/'
        if (not os.path.exists(data_path)):
            os.makedirs(data_path)
        for i, (img, label) in enumerate(zip(test_set[0], test_set[1])):
            img_path = data_path + str(i) + '.jpg'
            print('test_img_path:', img_path, 'test_img_num:', i)
            io.imsave(img_path, img.numpy())
            f.write(str(label.item()) + '\n')
        f.close()



convert_to_img(True)
convert_to_img(False)

按标签分别存储

按标签存储的意思就是，分别将图片按照标注的标签存放到从0-9的文件夹中，不需要手动建立文件夹。

import os
import torch
import torchvision
import torchvision.transforms as transforms
from skimage import io
import torchvision.datasets.mnist as mnist
import numpy

# Device configuration
device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')

# MNIST dataset
train_dataset = torchvision.datasets.MNIST(root='data/',
                                           train=True,
                                           transform=transforms.ToTensor(),
                                           download=True)

test_dataset = torchvision.datasets.MNIST(root='data/',
                                          train=False,
                                          transform=transforms.ToTensor())

root='./data/MNIST/raw/'

train_set=(mnist.read_image_file(os.path.join(root,'train-images-idx3-ubyte')),
           mnist.read_label_file(os.path.join(root,'train-labels-idx1-ubyte'))
           )
test_set = (
    mnist.read_image_file(os.path.join(root,'t10k-images-idx3-ubyte')),
    mnist.read_label_file(os.path.join(root,'t10k-labels-idx1-ubyte'))
)

def convert_to_img(train=True):
    if(train):
	        data_path='./data/data_train01/'
	
	        for i,(img,label) in enumerate(zip(train_set[0],train_set[1])):
	            img_path=data_path+str(label.item())+'/'
	            print('train_img_path:',img_path,'img_num:',i)
	            img_name=img_path+str(i)+'.png'
	            if (not os.path.exists(img_path)):
	                os.makedirs(img_path)
	            io.imsave(img_name, img.numpy())
    else:
		    data_path = './data/data_test01/'
		
		        for i, (img, label) in enumerate(zip(test_set[0], test_set[1])):
		            img_path = data_path +str(label.item())+ '/'
		            print('test_img_path:', img_path, 'img_numpy:',i)
		            img_name = img_path + str(i) + '.png'
		            if (not os.path.exists(img_path)):
		                os.makedirs(img_path)
		            io.imsave(img_name, img.numpy())

convert_to_img(True)
convert_to_img(False)

好的，代码部分结束，完成！