TensorFlow神经网络创建多层感知机MNIST数据集_Python

前面使用TensorFlow实现一个完整的Softmax Regression，并在MNIST数据及上取得了约92%的正确率。

前文传送门： TensorFlow教程Softmax逻辑回归识别手写数字MNIST数据集

现在建含一个隐层的神经网络模型（多层感知机）。

				?

									import tensorflow as tf

									import numpy as np

									import input_data

									mnist = input_data.read_data_sets('data/', one_hot=True)

									n_hidden_1 = 256

									n_input    = 784

									n_classes  = 10

									# INPUTS AND OUTPUTS

									x = tf.placeholder(tf.float32, [None, n_input]) # 用placeholder先占地方，样本个数不确定为None

									y = tf.placeholder(tf.float32, [None, n_classes]) # 用placeholder先占地方，样本个数不确定为None

									# NETWORK PARAMETERS

									weights = {

									    'w1': tf.Variable(tf.random_normal([n_input, n_hidden_1], stddev=0.1)),

									    'out': tf.Variable(tf.zeros([n_hidden_1, n_classes]))

									}

									biases = {

									    'b1': tf.Variable(tf.zeros([n_hidden_1])),

									    'out': tf.Variable(tf.zeros([n_classes]))

									}

									print("NETWORK READY")

									def multilayer_perceptron(_X, _weights, _biases): # 前向传播，l1、l2每一层后面加relu激活函数

									    layer_1 = tf.nn.relu(tf.add(tf.matmul(_X, _weights['w1']), _biases['b1'])) # 隐层

									    return (tf.matmul(layer_1, _weights['out']) + _biases['out']) # 返回输出层的结果，得到十个类别的得分值

									pred = multilayer_perceptron(x, weights, biases) # 前向传播的预测值

									cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(pred, y)) # 交叉熵损失函数，参数分别为预测值pred和实际label值y，reduce_mean为求平均loss

									optm = tf.train.GradientDescentOptimizer(0.01).minimize(cost) # 梯度下降优化器

									corr = tf.equal(tf.argmax(pred, 1), tf.argmax(y, 1)) # tf.equal()对比预测值的索引和实际label的索引是否一样，一样返回True，不一样返回False

									accr = tf.reduce_mean(tf.cast(corr, tf.float32)) # 将pred即True或False转换为1或0,并对所有的判断结果求均值

									init = tf.global_variables_initializer()

									print("FUNCTIONS READY")

									# 上面神经网络结构定义好之后，下面定义一些超参数

									training_epochs = 100 # 所有样本迭代100次

									batch_size = 100 # 每进行一次迭代选择100个样本

									display_step = 5

									# LAUNCH THE GRAPH

									sess = tf.Session() # 定义一个Session

									sess.run(init) # 在sess里run一下初始化操作

									# OPTIMIZE

									for epoch in range(training_epochs):

									    avg_cost = 0.

									    total_batch = int(mnist.train.num_examples/batch_size)

									    # Loop over all batches

									    for i in range(total_batch):

									        batch_xs, batch_ys = mnist.train.next_batch(batch_size) # 逐个batch的去取数据

									        sess.run(optm, feed_dict={x: batch_xs, y: batch_ys})

									        avg_cost += sess.run(cost, feed_dict={x: batch_xs, y: batch_ys})/total_batch

									    # Display logs per epoch step

									    if epoch % display_step == 0:

									        train_acc = sess.run(accr, feed_dict={x: batch_xs, y: batch_ys})

									        test_acc = sess.run(accr, feed_dict={x: mnist.test.images, y: mnist.test.labels})

									        print("Epoch: %03d/%03d cost: %.9f TRAIN ACCURACY: %.3f TEST ACCURACY: %.3f"

									              % (epoch, training_epochs, avg_cost, train_acc, test_acc))

									print("DONE")