MNIST Data Generator

mnist_data_gen.py

from tensorflow.examples.tutorials.mnist import input_data
import pandas as pd
import numpy as np

def convert_tensorflow_mnist_train_data_to_csv(path_to_save:str):
    mnist = input_data.read_data_sets("/tmp/data/", one_hot=True)
    data = pd.DataFrame(mnist.train.images)
    label = pd.DataFrame(mnist.train.labels)
    df = pd.concat([data, label], axis=1)
    df.columns = ['D_{}'.format(i) for i in range(784)] + ['L_{}'.format(i) for i in range(10)]
    print('Saving csv file to {}'.format(path_to_save))
    df.to_csv(path_to_save, index=False)

def convert_tensorflow_mnist_test_data_to_csv(path_to_save:str):
    mnist = input_data.read_data_sets("/tmp/data/", one_hot=True)
    data = pd.DataFrame(mnist.test.images)
    label = pd.DataFrame(mnist.test.labels)
    df = pd.concat([data, label], axis=1)
    df.columns = ['D_{}'.format(i) for i in range(784)] + ['L_{}'.format(i) for i in range(10)]
    print('Saving csv file to {}'.format(path_to_save))
    df.to_csv(path_to_save, index=False)


def random_data(dims, path_to_save):
    df = pd.DataFrame(np.random.randint(0,2, dims))
    df.columns = ['D_{}'.format(i) for i in range(dims[-1])]
    print('Saving csv file to {}'.format(path_to_save))
    df.to_csv(path_to_save, index=False)

convert_tensorflow_mnist_train_data_to_csv('/tmp/mnist_train.csv')
convert_tensorflow_mnist_test_data_to_csv('/tmp/mnist_test.csv')
random_data([input_data.read_data_sets("/tmp/data/", one_hot=True).train.num_examples, 200], '/tmp/random.csv')