TensorFlow入门（二）简单前馈网络实现 mnist 分类

在本教程中，我们来实现一个非常简单的两层全连接网络来完成MNIST数据的分类问题。 
输入[-1,28*28], FC1 有 1024 个neurons， FC2 有 10 个neurons。这么简单的一个全连接网络，结果测试准确率达到了 0.98。还是非常棒的！！！

import numpy as npimport tensorflow as tf# 设置按需使用GPUconfig = tf.ConfigProto()config.gpu_options.allow_growth = Truesess = tf.InteractiveSession(config=config)

1. 导入数据

# 用tensorflow 导入数据from tensorflow.examples.tutorials.mnist import input_datamnist = input_data.read_data_sets('MNIST_data', one_hot=True)

Extracting MNIST_data/train-images-idx3-ubyte.gzExtracting MNIST_data/train-labels-idx1-ubyte.gzExtracting MNIST_data/t10k-images-idx3-ubyte.gzExtracting MNIST_data/t10k-labels-idx1-ubyte.gz

print 'training data shape ', mnist.train.images.shapeprint 'training label shape ', mnist.train.labels.shape

training data shape  (55000, 784)training label shape  (55000, 10)

2. 构建网络

# 权值初始化def weight_variable(shape):    # 用正态分布来初始化权值    initial = tf.truncated_normal(shape, stddev=0.1)    return tf.Variable(initial)def bias_variable(shape):    # 本例中用relu激活函数，所以用一个很小的正偏置较好    initial = tf.constant(0.1, shape=shape)    return tf.Variable(initial)# input_layerX_ = tf.placeholder(tf.float32, [None, 784])y_ = tf.placeholder(tf.float32, [None, 10])# FC1W_fc1 = weight_variable([784, 1024])b_fc1 = bias_variable([1024])h_fc1 = tf.nn.relu(tf.matmul(X_, W_fc1) + b_fc1)# FC2W_fc2 = weight_variable([1024, 10])b_fc2 = bias_variable([10])y_pre = tf.nn.softmax(tf.matmul(h_fc1, W_fc2) + b_fc2)

3. 训练和评估

# 1.损失函数：cross_entropycross_entropy = -tf.reduce_sum(y_ * tf.log(y_pre))# 2.优化函数：AdamOptimizer, 优化速度要比 GradientOptimizer 快很多train_step = tf.train.AdamOptimizer(0.001).minimize(cross_entropy)# 3.预测结果评估#　预测值中最大值（１）即分类结果，是否等于原始标签中的（１）的位置。argmax()取最大值所在的下标correct_prediction = tf.equal(tf.argmax(y_pre, 1), tf.arg_max(y_, 1))  accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))# 开始运行sess.run(tf.global_variables_initializer())# 这大概迭代了不到 10 个 epoch， 训练准确率已经达到了0.98for i in range(5000):    X_batch, y_batch = mnist.train.next_batch(batch_size=100)    train_step.run(feed_dict={X_: X_batch, y_: y_batch})    if (i+1) % 200 == 0:        train_accuracy = accuracy.eval(feed_dict={X_: mnist.train.images, y_: mnist.train.labels})        print "step %d, training acc %g" % (i+1, train_accuracy)    if (i+1) % 1000 == 0:        test_accuracy = accuracy.eval(feed_dict={X_: mnist.test.images, y_: mnist.test.labels})        print "= " * 10, "step %d, testing acc %g" % (i+1, test_accuracy)

step 200, training acc 0.937364step 400, training acc 0.965818step 600, training acc 0.973364step 800, training acc 0.977709step 1000, training acc 0.981528= = = = = = = = = =  step 1000, testing acc 0.9688step 1200, training acc 0.988437step 1400, training acc 0.988728step 1600, training acc 0.987491step 1800, training acc 0.993873step 2000, training acc 0.992527= = = = = = = = = =  step 2000, testing acc 0.9789step 2200, training acc 0.995309step 2400, training acc 0.995455step 2600, training acc 0.9952step 2800, training acc 0.996073step 3000, training acc 0.9964= = = = = = = = = =  step 3000, testing acc 0.9778step 3200, training acc 0.996709step 3400, training acc 0.998109step 3600, training acc 0.997455step 3800, training acc 0.995055step 4000, training acc 0.997291= = = = = = = = = =  step 4000, testing acc 0.9808step 4200, training acc 0.997746step 4400, training acc 0.996073step 4600, training acc 0.998564step 4800, training acc 0.997946step 5000, training acc 0.998673= = = = = = = = = =  step 5000, testing acc 0.98