Hands-On Python Deep Learning for the Web

Creating a neural network using Keras

Now, we are ready to proceed with the creation of the neural network:

1. We will first create a Sequential neural network model in Keras:

model = Sequential()

2. To add a neuron layer to the network, we use the following code:

model.add(Conv2D(32, (3, 3), input_shape=(28,28,1)))

This adds a two-dimensional convolutional neuron layer to the network with an input shape that is the same as the shape of the images.

3. Now, let's add the activation layer with relu as the activation function:

model.add(Activation('relu'))

4. After adding the activation layer, we can perform a batch normalization. During training, the data passes through several computational layers and may become too large or too small. This is ...