Build ANN fully connected network with TensorFlow

""" Neural Network.

A 2-Hidden Layers Fully Connected Neural Network (a.k.a Multilayer Perceptron)
implementation with TensorFlow. This example is using the MNIST database
of handwritten digits (http://yann.lecun.com/exdb/mnist/).

This example is using TensorFlow layers, see 'neural_network_raw' example for
a raw implementation with variables.

Links:
    [MNIST Dataset](http://yann.lecun.com/exdb/mnist/).

Author: Aymeric Damien
Project: https://github.com/aymericdamien/TensorFlow-Examples/
"""

from __future__ import print_function

# Import MNIST data
from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets("/tmp/data/", one_hot=False)

import tensorflow as tf

# Parameters
learning_rate = 0.1
num_steps = 1000
batch_size = 128
display_step = 100

# Network Parameters
n_hidden_1 = 256 # 1st layer number of neurons
n_hidden_2 = 256 # 2nd layer number of neurons
num_input = 784 # MNIST data input (img shape: 28*28)
num_classes = 10 # MNIST total classes (0-9 digits)


# Define the neural network
def neural_net(x_dict):
    # TF Estimator input is a dict, in case of multiple inputs
    x = x_dict['images']
    # Hidden fully connected layer with 256 neurons
    layer_1 = tf.layers.dense(x, n_hidden_1)
    # Hidden fully connected layer with 256 neurons
    layer_2 = tf.layers.dense(layer_1, n_hidden_2)
    # Output fully connected layer with a neuron for each class
    out_layer = tf.layers.dense(layer_2, num_classes)
    return out_layer


# Define the model function (following TF Estimator Template)
def model_fn(features, labels, mode):
    # Build the neural network
    logits = neural_net(features)

    # Predictions
    pred_classes = tf.argmax(logits, axis=1)
    pred_probas = tf.nn.softmax(logits)

    # If prediction mode, early return
    if mode == tf.estimator.ModeKeys.PREDICT:
        return tf.estimator.EstimatorSpec(mode, predictions=pred_classes)

        # Define loss and optimizer
    loss_op = tf.reduce_mean(tf.nn.sparse_softmax_cross_entropy_with_logits(
        logits=logits, labels=tf.cast(labels, dtype=tf.int32)))
    optimizer = tf.train.GradientDescentOptimizer(learning_rate=learning_rate)
    train_op = optimizer.minimize(loss_op,
                                  global_step=tf.train.get_global_step())

    # Evaluate the accuracy of the model
    acc_op = tf.metrics.accuracy(labels=labels, predictions=pred_classes)

    # TF Estimators requires to return a EstimatorSpec, that specify
    # the different ops for training, evaluating, ...
    estim_specs = tf.estimator.EstimatorSpec (
        mode=mode,
        predictions=pred_classes,
        loss=loss_op,
        train_on=train_on,
        eval_metric_ops={'accuracy': acc_op})

    return estim_specs

# Build the Estimator
model = tf.estimator.Estimator(model_fn)

# Define the input function for training
input_fn = tf.estimator.inputs.numpy_input_fn(
    x={'images': mnist.train.images}, y=mnist.train.labels,
    batch_size=batch_size, num_epochs=None, shuffle=True)
# Train the Model
model.train(input_fn, steps=num_steps)

# Evaluate the Model
# Define the input function for evaluating
input_fn = tf.estimator.inputs.numpy_input_fn(
    x={'images': mnist.test.images}, y=mnist.test.labels,
    batch_size=batch_size, shuffle=False)
# Use the Estimator 'evaluate' method
e = model.evaluate(input_fn)

print("Testing Accuracy:", e['accuracy'])

# Predict single images
n_images = 4
# Get images from test set
test_images = mnist.test.images[:n_images]
# Prepare the input data
input_fn = tf.estimator.inputs.numpy_input_fn(
    x={'images': test_images}, shuffle=False)
# Use the model to predict the images class
preds = list(model.predict(input_fn))

# Display
for i in range(n_images):
    plt.imshow(np.reshape(test_images[i], [28, 28]), cmap='gray')
    plt.show()
    print("Model prediction:", preds[i])
    

The above example code uses the TensorFlow framework and the handwritten digit recognition dataset MNIST to train the network, and the running environment is Python3.6 and TensorFlow1.7. The neural network construction process is summarized as follows:

step1. Hyperparameter setting. Such as the learning rate, the number of training steps, the mini-batch size, and the interval step size for printing out the training results.

         Neural network parameter settings. Such as the number of neural units in the input layer, hidden layer, and output layer.

step2. Define the neural network forward propagation process function.

step3. Define the model_fn function. This function returns an estimator with some parameters specified. See the TensorFlow API for function definitions.

step4. Define the model. The model is built by tf.estimator.Estimator(model_fn).

step5. Define the input_fn function. See the TensorFlow API for function definitions.

step6. Train the model.

step7. Define the input_fn function.

step8. Evaluate the model.