Tensorflow学习教程------普通神经网络对mnist数据集分类
首先是不含隐层的神经网络, 输入层是784个神经元 输出层是10个神经元
代码如下
#coding:utf-8 import tensorflow as tf from tensorflow.examples.tutorials.mnist import input_data #载入数据集 mnist = input_data.read_data_sets("MNIST_data", one_hot=True) #每个批次的大小 batch_size = 100 #计算一共有多少个批次 n_batch = mnist.train.num_examples // batch_size #定义两个placeholder x = tf.placeholder(tf.float32, [None,784]) #输入图像 y = tf.placeholder(tf.float32, [None,10]) #输入标签 #创建一个简单的神经网络 784个像素点对应784个数 因此输入层是784个神经元 输出层是10个神经元 不含隐层 #最后准确率在92%左右 W = tf.Variable(tf.zeros([784,10])) #生成784行 10列的全0矩阵 b = tf.Variable(tf.zeros([1,10])) prediction = tf.nn.softmax(tf.matmul(x,W)+b) #二次代价函数 loss = tf.reduce_mean(tf.square(y-prediction)) #使用梯度下降法 train_step = tf.train.GradientDescentOptimizer(0.2).minimize(loss) #初始化变量 init = tf.global_variables_initializer() #结果存放在布尔型列表中 #argmax能给出某个tensor对象在某一维上的其数据最大值所在的索引值 correct_prediction = tf.equal(tf.argmax(y,1), tf.argmax(prediction,1)) #求准确率 accuracy = tf.reduce_mean(tf.cast(correct_prediction,tf.float32)) with tf.Session() as sess: sess.run(init) for epoch in range(21): #21个epoch 把所有的图片训练21次 for batch in range(n_batch): # batch_xs,batch_ys = mnist.train.next_batch(batch_size) sess.run(train_step,feed_dict={x:batch_xs,y:batch_ys}) acc = sess.run(accuracy,feed_dict={x:mnist.test.images, y:mnist.test.labels}) print ("Iter " + str(epoch) + ",Testing Accuracy " + str(acc))
结果如下
Iter 0,Testing Accuracy 0.8304 Iter 1,Testing Accuracy 0.8704 Iter 2,Testing Accuracy 0.8821 Iter 3,Testing Accuracy 0.8876 Iter 4,Testing Accuracy 0.8932 Iter 5,Testing Accuracy 0.8968 Iter 6,Testing Accuracy 0.8995 Iter 7,Testing Accuracy 0.9019 Iter 8,Testing Accuracy 0.9033 Iter 9,Testing Accuracy 0.9048 Iter 10,Testing Accuracy 0.9065 Iter 11,Testing Accuracy 0.9074 Iter 12,Testing Accuracy 0.9084 Iter 13,Testing Accuracy 0.909 Iter 14,Testing Accuracy 0.9094 Iter 15,Testing Accuracy 0.9112 Iter 16,Testing Accuracy 0.9117 Iter 17,Testing Accuracy 0.9128 Iter 18,Testing Accuracy 0.9127 Iter 19,Testing Accuracy 0.9132 Iter 20,Testing Accuracy 0.9144
接下来是含一个隐层的神经网络,输入层是784个神经元,两个隐层都是100个神经元,输出层是10个神经元,迭代500次,最后准确率在88%左右,汗。。。。准确率反而降低了,慢慢调参吧
#coding:utf-8 import tensorflow as tf from tensorflow.examples.tutorials.mnist import input_data #载入数据集 mnist = input_data.read_data_sets("MNIST_data", one_hot=True) #每个批次的大小 batch_size = 50 #计算一共有多少个批次 n_batch = mnist.train.num_examples // batch_size #定义两个placeholder x = tf.placeholder(tf.float32, [None,784]) #输入图像 y = tf.placeholder(tf.float32, [None,10]) #输入标签 #定义神经网络中间层 Weights_L1 = tf.Variable(tf.random_normal([784,100])) biase_L1 = tf.Variable(tf.zeros([1,100])) Wx_plus_b_L1 = tf.matmul(x, Weights_L1)+biase_L1 L1 = tf.nn.tanh(Wx_plus_b_L1) #使用正切函数作为激活函数 Weights_L2 = tf.Variable(tf.random_normal([100,100])) biase_L2 = tf.Variable(tf.zeros([1,100])) Wx_plus_b_L2 = tf.matmul(L1, Weights_L2)+biase_L2 L2 = tf.nn.tanh(Wx_plus_b_L2) #使用正切函数作为激活函数 #定义神经网络输出层 Weights_L3 = tf.Variable(tf.random_normal([100,10])) biase_L3 = tf.Variable(tf.zeros([1,10])) Wx_plus_b_L3 = tf.matmul(L2,Weights_L3) + biase_L3 prediction = tf.nn.tanh(Wx_plus_b_L3) #二次代价函数 loss = tf.reduce_mean(tf.square(y-prediction)) #使用梯度下降法 train_step = tf.train.GradientDescentOptimizer(0.2).minimize(loss) #初始化变量 init = tf.global_variables_initializer() #结果存放在布尔型列表中 #argmax能给出某个tensor对象在某一维上的其数据最大值所在的索引值 correct_prediction = tf.equal(tf.argmax(y,1), tf.argmax(prediction,1)) #求准确率 accuracy = tf.reduce_mean(tf.cast(correct_prediction,tf.float32)) with tf.Session() as sess: sess.run(init) for epoch in range(500): for batch in range(n_batch): batch_xs,batch_ys = mnist.train.next_batch(batch_size) sess.run(train_step,feed_dict={x:batch_xs,y:batch_ys}) acc = sess.run(accuracy,feed_dict={x:mnist.test.images, y:mnist.test.labels}) print ("Iter " + str(epoch) + ",Testing Accuracy " + str(acc))
Iter 487,Testing Accuracy 0.8847 Iter 488,Testing Accuracy 0.8853 Iter 489,Testing Accuracy 0.878 Iter 490,Testing Accuracy 0.8861 Iter 491,Testing Accuracy 0.8863 Iter 492,Testing Accuracy 0.8784 Iter 493,Testing Accuracy 0.8855 Iter 494,Testing Accuracy 0.8787 Iter 495,Testing Accuracy 0.881 Iter 496,Testing Accuracy 0.8837 Iter 497,Testing Accuracy 0.8817 Iter 498,Testing Accuracy 0.8837 Iter 499,Testing Accuracy 0.8866