python调用数据集mnist_pythonMNIST手写识别数据调用API的方法

MNIST数据集比较小&＃xff0c;一般入门机器学习都会采用这个数据集来训练

有4个有用的文件&＃xff1a;

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&＃xff0c;黑白图片。

下面给出load数据集的代码&＃xff1a;

import os

import struct

import numpy as np

import matplotlib.pyplot as plt

def load_mnist():

&＃39;&＃39;&＃39;

Load mnist data

http://yann.lecun.com/exdb/mnist/

60000 training examples

10000 test sets

Arguments:

kind: &＃39;train&＃39; or &＃39;test&＃39;, string charater input with a default value &＃39;train&＃39;

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)

&＃39;&＃39;&＃39;

root_path &＃61; &＃39;/home/cc/deep_learning/data_sets/mnist&＃39;

train_labels_path &＃61; os.path.join(root_path, &＃39;train-labels.idx1-ubyte&＃39;)

train_images_path &＃61; os.path.join(root_path, &＃39;train-images.idx3-ubyte&＃39;)

test_labels_path &＃61; os.path.join(root_path, &＃39;t10k-labels.idx1-ubyte&＃39;)

test_images_path &＃61; os.path.join(root_path, &＃39;t10k-images.idx3-ubyte&＃39;)

with open(train_labels_path, &＃39;rb&＃39;) as lpath:

# &＃39;>&＃39; denotes bigedian

# &＃39;I&＃39; denotes unsigned char

magic, n &＃61; struct.unpack(&＃39;>II&＃39;, lpath.read(8))

#loaded &＃61; np.fromfile(lpath, dtype &＃61; np.uint8)

train_labels &＃61; np.fromfile(lpath, dtype &＃61; np.uint8).astype(np.float)

with open(train_images_path, &＃39;rb&＃39;) as ipath:

magic, num, rows, cols &＃61; struct.unpack(&＃39;>IIII&＃39;, ipath.read(16))

loaded &＃61; np.fromfile(train_images_path, dtype &＃61; np.uint8)

# images start from the 16th bytes

train_images &＃61; loaded[16:].reshape(len(train_labels), 784).astype(np.float)

with open(test_labels_path, &＃39;rb&＃39;) as lpath:

# &＃39;>&＃39; denotes bigedian

# &＃39;I&＃39; denotes unsigned char

magic, n &＃61; struct.unpack(&＃39;>II&＃39;, lpath.read(8))

#loaded &＃61; np.fromfile(lpath, dtype &＃61; np.uint8)

test_labels &＃61; np.fromfile(lpath, dtype &＃61; np.uint8).astype(np.float)

with open(test_images_path, &＃39;rb&＃39;) as ipath:

magic, num, rows, cols &＃61; struct.unpack(&＃39;>IIII&＃39;, ipath.read(16))

loaded &＃61; np.fromfile(test_images_path, dtype &＃61; np.uint8)

# images start from the 16th bytes

test_images &＃61; 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&＃xff1a;

def test_mnist_data():

&＃39;&＃39;&＃39;

Just to check the data

Argument:

none

Return:

none

&＃39;&＃39;&＃39;

train_images, train_labels, test_images, test_labels &＃61; load_mnist()

fig, ax &＃61; plt.subplots(nrows &＃61; 2, ncols &＃61; 5, sharex &＃61; True, sharey &＃61; True)

ax &＃61;ax.flatten()

for i in range(10):

img &＃61; train_images[i][:].reshape(28, 28)

ax[i].imshow(img, cmap &＃61; &＃39;Greys&＃39;, interpolation &＃61; &＃39;nearest&＃39;)

print(&＃39;corresponding labels &＃61; %d&＃39; %train_labels[i])

if __name__ &＃61;&＃61; &＃39;__main__&＃39;:

test_mnist_data()

跑出的结果如下&＃xff1a;

以上就是本文的全部内容&＃xff0c;希望对大家的学习有所帮助&＃xff0c;也希望大家多多支持脚本之家。