First, load the MNIST dataset, and define the training and test features and the targets using the following code:
from tensorflow.examples.tutorials.mnist import input_datamnist_home = os.path.join(datasetslib.datasets_root, 'mnist')mnist = input_data.read_data_sets(mnist_home, one_hot=True)X_train = mnist.train.imagesX_test = mnist.test.imagesY_train = mnist.train.labelsY_test = mnist.test.labelsnum_outputs = 10 # 0-9 digitsnum_inputs = 784 # total pixels
We create three helper functions that will help us create a simple MLP with only one hidden layer, followed by a larger MLP with multiple layers and multiple neurons in each layer.
The mlp() function builds the network layers with the following ...
