利用全连接网络将图片分类

一 实例描述

构建一个简单的多层神经网络，以拟合MNIST样本特征完成分类任务。

二 代码

import tensorflow as tf
# 导入 MINST 数据集
from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets("/data/", one_hot=True)
#定义网络参数
'''
在输入层和输出层之间使用两个隐藏层，每层256个节点，学习率使用0.001。
'''
learning_rate = 0.001
training_epochs = 25
batch_size = 100
display_step = 1
# Network Parameters
n_hidden_1 = 256 # 1st layer number of features
n_hidden_2 = 256 # 2nd layer number of features
n_input = 784 # MNIST data 输入 (img shape: 28*28)
n_classes = 10  # MNIST 列别 (0-9 ，一共10类)
'''
定义网络结构
'''
# tf Graph input
x = tf.placeholder("float", [None, n_input])
y = tf.placeholder("float", [None, n_classes])
# Create model
#multilayer_perceptron函数封装好网络模型函数，第一层与第二层均使用Relu激活函数，loss使用交叉熵
def multilayer_perceptron(x, weights, biases):
    # Hidden layer with RELU activation
    layer_1 = tf.add(tf.matmul(x, weights['h1']), biases['b1'])
    layer_1 = tf.nn.relu(layer_1)
    # Hidden layer with RELU activation
    layer_2 = tf.add(tf.matmul(layer_1, weights['h2']), biases['b2'])
    layer_2 = tf.nn.relu(layer_2)
    # Output layer with linear activation
    out_layer = tf.matmul(layer_2, weights['out']) + biases['out']
    return out_layer
    
# Store layers weight & bias
weights = {
    'h1': tf.Variable(tf.random_normal([n_input, n_hidden_1])),
    'h2': tf.Variable(tf.random_normal([n_hidden_1, n_hidden_2])),
    'out': tf.Variable(tf.random_normal([n_hidden_2, n_classes]))
}
biases = {
    'b1': tf.Variable(tf.random_normal([n_hidden_1])),
    'b2': tf.Variable(tf.random_normal([n_hidden_2])),
    'out': tf.Variable(tf.random_normal([n_classes]))
}
# 构建模型
pred = multilayer_perceptron(x, weights, biases)
# Define loss and optimizer
cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=pred, labels=y))
optimizer = tf.train.AdamOptimizer(learning_rate=learning_rate).minimize(cost)
# 初始化变量
init = tf.global_variables_initializer()
# 启动session
with tf.Session() as sess:
    sess.run(init)
    # 启动循环开始训练
    for epoch in range(training_epochs):
        avg_cost = 0.
        total_batch = int(mnist.train.num_examples/batch_size)
        # 遍历全部数据集
        for i in range(total_batch):
            batch_x, batch_y = mnist.train.next_batch(batch_size)
            # Run optimization op (backprop) and cost op (to get loss value)
            _, c = sess.run([optimizer, cost], feed_dict={x: batch_x,
                                                          y: batch_y})
            # Compute average loss
            avg_cost += c / total_batch
        # 显示训练中的详细信息
        if epoch % display_step == 0:
            print ("Epoch:", '%04d' % (epoch+1), "cost=", \
                "{:.9f}".format(avg_cost))
    print (" Finished!")
    # 测试 model
    correct_prediction = tf.equal(tf.argmax(pred, 1), tf.argmax(y, 1))
    # 计算准确率
    accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float"))
    print ("Accuracy:", accuracy.eval({x: mnist.test.images, y: mnist.test.labels}))

三 运行结果

Successfully downloaded train-images-idx3-ubyte.gz 9912422 bytes.

Extracting /data/train-images-idx3-ubyte.gz

Successfully downloaded train-labels-idx1-ubyte.gz 28881 bytes.

Extracting /data/train-labels-idx1-ubyte.gz

Successfully downloaded t10k-images-idx3-ubyte.gz 1648877 bytes.

Extracting /data/t10k-images-idx3-ubyte.gz

Successfully downloaded t10k-labels-idx1-ubyte.gz 4542 bytes.

Extracting /data/t10k-labels-idx1-ubyte.gz

Epoch: 0001 cost= 167.503109946

Epoch: 0002 cost= 42.023845719

Epoch: 0003 cost= 26.434357550

Epoch: 0004 cost= 18.450763896

Epoch: 0005 cost= 13.571702020

Epoch: 0006 cost= 10.169877889

Epoch: 0007 cost= 7.683901399

Epoch: 0008 cost= 5.580870980

Epoch: 0009 cost= 4.258601876

Epoch: 0010 cost= 3.213997342

Epoch: 0011 cost= 2.429289338

Epoch: 0012 cost= 1.821290299

Epoch: 0013 cost= 1.474814914

Epoch: 0014 cost= 1.091195856

Epoch: 0015 cost= 0.901945515

Epoch: 0016 cost= 0.752033565

Epoch: 0017 cost= 0.625414731

Epoch: 0018 cost= 0.596724849

Epoch: 0019 cost= 0.513915204

Epoch: 0020 cost= 0.432392413

Epoch: 0021 cost= 0.511673744

Epoch: 0022 cost= 0.384512378

Epoch: 0023 cost= 0.323022437

Epoch: 0024 cost= 0.367242469

Epoch: 0025 cost= 0.299307836

Finished!

Accuracy: 0.9534

四 说明

全连接网络可以成功将图片进行分类，并且随着层数的增加和节点的增多，还能够得到更好的拟合效果。

五 注意

由于神经网络学习算法限制，在实际情况中并不是层数越多，节点越多，效果就越好，因为在训练过程中使用BP算法，会随着层数的逐渐增大其算出来的调整值会逐渐变小，直到其他层都感觉不到变化，即梯度消失的情况。