Tensorflow 2: use neural network to classify and compare fashion MNIST

Experimental setup

  In this experiment, neural networks with different activation functions are used to MNIST Data sets are classified , At the same time, the influence of the number of hidden layers on the experimental results . The hidden layer of the experiment is mainly 1 Layer and the 2 layer , The activation functions are ReLU、Laaky ReLU and Softmax, Output functions are unchanged Softmax. The experimental data set is Fashion MNIST Data sets , It is mainly used for multi classification of handwriting recognition .

Code and experimental results

Hidden layer is 1 The different activation functions of layer

# TensorFlow and tf.keras
import tensorflow as tf
from tensorflow import keras

# Helper libraries
import numpy as np

# Fashion MNIST input data
fashion_mnist = keras.datasets.fashion_mnist
(train_images, train_labels), (test_images, test_labels) = fashion_mnist.load_data()

# preprocess
train_images = train_images / 255.0
test_images = test_images / 255.0

# build model, loss, and train op
model = keras.Sequential([
    keras.layers.Flatten(input_shape=(28, 28)),
    keras.layers.Dense(196, activation='relu'), # change the activation
    keras.layers.Dense(10, activation='softmax')
])

model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy'])
model.fit(train_images, train_labels, epochs=10)

test_loss, test_acc = model.evaluate(test_images, test_labels)
print('\nTest Accuracy:', test_acc)

  relu

  elu

  softmax

Hidden layer is 2 Performance of different activation functions in layer

# build model, loss, and train op
model = keras.Sequential([
    keras.layers.Flatten(input_shape=(28, 28)),
    keras.layers.Dense(196, activation='relu'), # change activation
    keras.layers.Dense(98, activation='relu'), # change activation
    keras.layers.Dense(10, activation='softmax')
])

  relu

  elu

  softmax