import torch
import torchvision 
import torch.utils.data as Data 
import scipy.misc
import os
import matplotlib.pyplot as plt   
BATCH_SIZE = 50  
DOWNLOAD_MNIST = True 

train_data = torchvision.datasets.MNIST(root='./mnist/', train=True,transform=torchvision.transforms.ToTensor(),              
  download=DOWNLOAD_MNIST, )
test_data = torchvision.datasets.MNIST(root='./mnist/', train=False)

train_loader = Data.DataLoader(dataset=train_data, batch_size=BATCH_SIZE, shuffle=True)
test_x = Variable(torch.unsqueeze(test_data.test_data, dim=1), requires_grad=True).type(torch.FloatTensor)[:20]/255 
test_y = test_data.test_labels[:20]#前两千张
 #具体查看图像形式为：
 
a_data, a_label = train_data[0]
print(type(a_data))#tensor 类型
#print(a_data)
print(a_label)

#把原始图片保存至MNIST_data/raw/下
save_dir="mnist/raw/"
if os.path.exists(save_dir)is False:
 os.makedirs(save_dir)
 
for i in range(20):
 image_array,_=train_data[i]#打印第i个
 image_array=image_array.resize(28,28)
 filename=save_dir + 'mnist_train_%d.jpg' % i#保存文件的格式
 print(filename)
 print(train_data.train_labels[i])#打印出标签
 scipy.misc.toimage(image_array,cmin=0.0,cmax=1.0).save(filename)#保存图像