多层神经网络的构建与实现

载入数据

import tensorflow as tf
import tensorflow.examples.tutorials.mnist.input_data as input_data

mnist = input_data.read_data_sets("/Users/liuqi/Desktop/data/MNIST_data/", one_hot=True)

---

构建模型

定义全连接层函数

# inputs:输入数据
# input_dim: 输入神经元数量
# output_dim: 输出神经元数量
# activation: 激活函数

def function_layer(inputs, input_dim, output_dim, activation = None):
    # 以截断正态分布的随机数初始化W
    W = tf.Variable(tf.truncated_normal([input_dim, output_dim], stddev = 0.1))
    
    # 以0初始化b
    b = tf.Variable(tf.zeros([output_dim]))
    
    XWb = tf.matmul(inputs, W) + b
    
    # 若无激活函数 则XWb直接为输出结果
    if activation is None:
        outputs = XWb
    else:
        # 若有激活函数  则对其输出结果进行变换
        outputs = activation(XWb)
    
    return outputs

构建输入层

x = tf.placeholder(tf.float32, [None, 784], name = "X")
y = tf.placeholder(tf.float32, [None, 10], name = "Y")

构建隐藏层

# 隐藏层1
h1 = function_layer(inputs=x, input_dim=784, output_dim=H1_NN, activation=tf.nn.relu)

# 隐藏层2
h2 = function_layer(inputs=h1, input_dim=H1_NN, output_dim=H2_NN, activation=tf.nn.relu)

# 隐藏层3
h3 = function_layer(inputs=h2, input_dim=H2_NN, output_dim=H3_NN, activation=tf.nn.relu)

构建输出层

forward = function_layer(inputs=h3, input_dim=H3_NN, output_dim=10, activation=None)
pred = tf.nn.softmax(forward)

---

训练模型

定义损失函数

loss_function = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=forward, labels = y))

设置训练参数

train_epoch = 40
batch_size = 50
total_batch = int(mnist.train.num_examples / batch_size)
display_step = 1
learning_rate = 0.01

选择优化器

optimizer = tf.train.AdamOptimizer(learning_rate).minimize(loss_function)

定义准确率

correct_predicton = tf.equal(tf.argmax(y, 1), tf.argmax(pred, 1))
accuracy = tf.reduce_mean(tf.cast(correct_predicton, tf.float32))

开始训练

# 记录训练开始时间
from time import time
startTime = time()

sess = tf.Session()
sess.run(tf.global_variables_initializer())

for epoch in range(train_epoch):
    for batch in range(total_batch):
        xs, ys = mnist.train.next_batch(batch_size)
        sess.run(optimizer, feed_dict={x:xs, y:ys})
    
    # 一轮训练完后，使用验证数据计算误差与准确率
    loss, acc = sess.run([loss_function, accuracy], feed_dict={x:mnist.validation.images, y:mnist.validation.labels})
    
    if (epoch+1) % display_step == 0:
        print("Train Epoch:", "%02d" % (epoch+1), "Loss=", "{:.9f}".format(loss), "Accuracy=", "{:.4f}".format(acc))
        
# 显示运行总时间
duration = time() - startTime
print("Train finished takes:", "{:.2f}".format(duration))

---

评估模型

acc_test = sess.run(accuracy, feed_dict={x:mnist.test.images, y:mnist.test.labels})
print("Test Accuracy:", acc_test)

参考视频：深度学习应用开发TensorFlow实践