python MNIST手写识别数据调用API的方法

MNIST数据集比较小，一般入门机器学习都会采用这个数据集来训练

下载地址：yann.lecun.com/exdb/mnist/

有4个有用的文件：
train-images-idx3-ubyte: training set images
train-labels-idx1-ubyte: training set labels
t10k-images-idx3-ubyte: test set images
t10k-labels-idx1-ubyte: test set labels

The training set contains 60000 examples, and the test set 10000 examples. 数据集存储是用binary file存储的，黑白图片。

下面给出load数据集的代码：

import os
import struct
import numpy as np
import matplotlib.pyplot as plt

def load_mnist():
 '''
 Load mnist data
 http://yann.lecun.com/exdb/mnist/

60000 training examples
 10000 test sets

Arguments:
   kind: 'train' or 'test', string charater input with a default value 'train'

Return:
   xxx_images: n*m array, n is the sample count, m is the feature number which is 28*28
   xxx_labels: class labels for each image, (0-9)
 '''

root_path = '/home/cc/deep_learning/data_sets/mnist'

train_labels_path = os.path.join(root_path, 'train-labels.idx1-ubyte')
 train_images_path = os.path.join(root_path, 'train-images.idx3-ubyte')

test_labels_path = os.path.join(root_path, 't10k-labels.idx1-ubyte')
 test_images_path = os.path.join(root_path, 't10k-images.idx3-ubyte')

with open(train_labels_path, 'rb') as lpath:
   # '>' denotes bigedian
   # 'I' denotes unsigned char
   magic, n = struct.unpack('>II', lpath.read(8))
   #loaded = np.fromfile(lpath, dtype = np.uint8)
   train_labels = np.fromfile(lpath, dtype = np.uint8).astype(np.float)

with open(train_images_path, 'rb') as ipath:
   magic, num, rows, cols = struct.unpack('>IIII', ipath.read(16))
   loaded = np.fromfile(train_images_path, dtype = np.uint8)
   # images start from the 16th bytes
   train_images = loaded[16:].reshape(len(train_labels), 784).astype(np.float)

with open(test_labels_path, 'rb') as lpath:
   # '>' denotes bigedian
   # 'I' denotes unsigned char
   magic, n = struct.unpack('>II', lpath.read(8))
   #loaded = np.fromfile(lpath, dtype = np.uint8)
   test_labels = np.fromfile(lpath, dtype = np.uint8).astype(np.float)

with open(test_images_path, 'rb') as ipath:
   magic, num, rows, cols = struct.unpack('>IIII', ipath.read(16))
   loaded = np.fromfile(test_images_path, dtype = np.uint8)
   # images start from the 16th bytes
   test_images = loaded[16:].reshape(len(test_labels), 784)  

return train_images, train_labels, test_images, test_labels

再看看图片集是什么样的：

def test_mnist_data():
 '''
 Just to check the data

Argument:
   none

Return:
   none
 '''
 train_images, train_labels, test_images, test_labels = load_mnist()
 fig, ax = plt.subplots(nrows = 2, ncols = 5, sharex = True, sharey = True)
 ax =ax.flatten()
 for i in range(10):
   img = train_images[i][:].reshape(28, 28)
   ax[i].imshow(img, cmap = 'Greys', interpolation = 'nearest')
   print('corresponding labels = ％d' ％train_labels[i])

if __name__ == '__main__':
 test_mnist_data()

跑出的结果如下：

来源：https://blog.csdn.net/caichao08/article/details/78988389