Python PyTorch 获取 MNIST 数据

1 PyTorch 获取 MNIST 数据

import torch
import numpy as np
import matplotlib.pyplot as plt # type: ignore
from torchvision import datasets, transforms

def mnist_get():
    print(torch.__version__)
    # 定义数据转换
    transform = transforms.Compose([
        transforms.ToTensor(),  # 将图像转换为张量
        transforms.Normalize((0.5,), (0.5,))  # 归一化图像数据
    ])
    # 获取数据
    train_data = datasets.MNIST(root='./data', train=False, download=True, transform=transform)
    test_data = datasets.MNIST(root='./data', train=False, download=True, transform=transform)
    # 训练数据
    train_image = train_data.data.numpy()
    train_label = train_data.targets.numpy()
    # 测试数据
    test_image = test_data.data.numpy()
    test_label = test_data.targets.numpy()
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21

2 PyTorch 保存 MNIST 数据

import torch
import numpy as np
import matplotlib.pyplot as plt # type: ignore
from torchvision import datasets, transforms

def mnist_save(mnist_path):
    print(torch.__version__)
    # 定义数据转换
    transform = transforms.Compose([
        transforms.ToTensor(),  # 将图像转换为张量
        transforms.Normalize((0.5,), (0.5,))  # 归一化图像数据
    ])
    # 获取数据
    train_data = datasets.MNIST(root='./data', train=False, download=True, transform=transform)
    test_data = datasets.MNIST(root='./data', train=False, download=True, transform=transform)
    # 训练数据
    train_image = train_data.data.numpy()
    train_label = train_data.targets.numpy()
    # 测试数据
    test_image = test_data.data.numpy()
    test_label = test_data.targets.numpy()
    np.savez(mnist_path, train_data=train_image, train_label=train_label, test_data=test_image, test_label=test_label)

mnist_path = 'C:\\Users\\Hyacinth\\Desktop\\mnist.npz'
mnist_save(mnist_path)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

3 PyTorch 显示 MNIST 数据

import torch
import numpy as np
import matplotlib.pyplot as plt # type: ignore
from torchvision import datasets, transforms

def mnist_show(mnist_path):
    data = np.load(mnist_path)
    image = data['train_data'][0:100]
    label = data['train_label'].reshape(-1, )
    plt.figure(figsize = (10, 10))
    for i in range(100):
        print('%f, %f' % (i, label[i]))
        plt.subplot(10, 10, i + 1)
        plt.imshow(image[i])
    plt.show()

mnist_path = 'C:\\Users\\Hyacinth\\Desktop\\mnist.npz'
mnist_show(mnist_path)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18