MNIST for Numpy

The MNIST database of handwritten digits has 60,000 training examples, and 10,000 test examples. Each example included in the MNIST database is a 28x28 grayscale image of handwritten digit and its corresponding label(0-9). This Python module makes it easy to load the MNIST database into numpy arrays. For more details about the MNIST database, please visit here.

Requirements

• Python 3.x
• Numpy

Usage

First, download mnist.py from this repository and locate it to your working directory. Then you can make your MNIST data to be prepared in Python as follows.

import mnist

mnist.init()

init() consists of 2-steps.

1. Download the MNIST database.
2. Make it into numpy arrays and save it as Pickle. (mnist.pkl)

You can do this in command line tools as well.

$ python3 mnist.py

After preparing, you can load the MNIST database into numpy array as follows.

x_train, t_train, x_test, t_test = mnist.load()

load() takes mnist.pkl and returns 4 numpy arrays.

• x_train : 60,000x784 numpy array that each row contains flattened version of training images.
• t_train : 1x60,000 numpy array that each component is true label of the corresponding training images.
• x_test : 10,000x784 numpy array that each row contains flattened version of test images.
• t_test : 1x10,000 numpy array that each component is true label of the corresponding test images.

Visualization

If you want to visualize each image in x_train and x_test, you can use the following methods.

For PIL (Python Imaging Library or Pillow) :

from PIL import Image

img = Image.fromarray(x_train[0,:].reshape(28,28)) # First image in the training set.
img.show() # Show the image

For Matplotlib :

import matplotlib.pyplot as plt

img = x_train[0,:].reshape(28,28) # First image in the training set.
plt.imshow(img,cmap='gray')
plt.show() # Show the image

Notice

Once you get mnist.pkl, you don't need to call init() anymore. Everything you need to do is to locate mnist.py and mnist.pkl in your working directory and to call load().