''' Created on Apr 21, 2017 @author: P0079482 ''' #-*- coding:utf-8 -*- import tensorflow as tf #定义神经网络结构相关的参数 INPUT_NODE=784 OUTPUT_NODE=10 LAYER1_NODE=500 #通过tf.get_variable函数来获取变量。在训练神经网络时会创建这些变量： #在测试时会通过保存的模型加载这些变量的取值。而且更加方便的是，因为可以在变量加载时 #将滑动平均变量重命名，所以可以直接通过同样的名字在训练时使用变量自身， #而在测试时使用变量的滑动平均值。在这个函数中也会将变量的正则化损失加入损失集合 def get_weight_variable(shape,regularizer): weights =tf.get_variable("weights",shape,initializer=tf.truncated_normal_initializer(stddev=0.1)) #当给出了正则化生成函数时，将当前变量的正则化损失加入名字为losses的集合。 #在这里使用了add_to_collection函数将一个张量加入一个集合，而这个集合的名称为losses #这是自定义的集合，不在Tensorflow自动管理的集合列表中 if regularizer!=None: tf.add_to_collection('losses',regularizer(weights)) return weights #定义神经网络的前向传播过程 def inference(input_tensor,regularizer): #声明第一层神经网络的变量并完成前向传播过程 with tf.variable_scope('layer1'): #这里通过tf.get_variable或tf.Variable没有本质区别， #因为在训练或是测试中没有在同一个程序中多次调用这个函数，如果在同一个程序中多次调用，在第一次调用之后 #需要将reuse参数设置为True weights=get_weight_variable([INPUT_NODE,LAYER1_NODE], regularizer) biases = tf.get_variable("biases", [LAYER1_NODE], initializer=tf.constant_initializer(0.0)) layer1=tf.nn.relu(tf.matmul(input_tensor,weights)+biases) #类似地声明第二层神经网络的变量并完成前向传播过程 with tf.variable_scope('layer2'): weights=get_weight_variable([LAYER1_NODE,OUTPUT_NODE], regularizer) biases=tf.get_variable("biases",[OUTPUT_NODE],initializer=tf.constant_initializer(0.0)) layer2=tf.matmul(layer1,weights)+biases #返回最后前向传播的结果 return layer2 #在上面这段代码中定义了神经网络的前向传播算法。无论是训练还是测试时 #都可以直接调用inference这个函数，而不用关心具体的神经网络结构 #使用定义好的前行传播过程，以下代码给出了神经网络的训练程序mnist_train.py
上面的代码是mnist_inference.py。里边定义了网络的前向传播结果的计算，封装了权重的二范式的计算
''' Created on 2017年4月21日 @author: weizhen ''' import os import tensorflow as tf from tensorflow.examples.tutorials.mnist import input_data # 加载mnist_inference.py中定义的常量和前向传播的函数 import mnist_inference # 配置神经网络的参数 BATCH_SIZE = 100 LEARNING_RATE_BASE = 0.8 LEARNING_RATE_DECAY = 0.99 REGULARAZTION_RATE = 0.0001 TRAINING_STEPS = 30000 MOVING_AVERAGE_DECAY = 0.99 # 模型保存的路径和文件名 MODEL_SAVE_PATH = "/path/to/model/" MODEL_NAME = "model.ckpt" def train(mnist): # 定义输入输出placeholder x = tf.placeholder(tf.float32, [None, mnist_inference.INPUT_NODE], name='x-input') y_ = tf.placeholder(tf.float32, [None, mnist_inference.OUTPUT_NODE], name='y-input') regularizer = tf.contrib.layers.l2_regularizer(REGULARAZTION_RATE) # 直接使用mnist_inference.py中定义的前向传播过程 y = mnist_inference.inference(x, regularizer) global_step = tf.Variable(0, trainable=False) # 和5.2.1小节样例中类似地定义损失函数、学习率、滑动平均操作以及训练过程 variable_averages = tf.train.ExponentialMovingAverage(MOVING_AVERAGE_DECAY, global_step) variable_averages_op = variable_averages.apply(tf.trainable_variables()) cross_entropy = tf.nn.sparse_softmax_cross_entropy_with_logits(logits=y, labels=tf.argmax(y_, 1)) cross_entropy_mean = tf.reduce_mean(cross_entropy) loss = cross_entropy_mean + tf.add_n(tf.get_collection('losses')) learning_rate = tf.train.exponential_decay(LEARNING_RATE_BASE, global_step, mnist.train.num_examples / BATCH_SIZE, LEARNING_RATE_DECAY) train_step = tf.train.GradientDescentOptimizer(learning_rate).minimize(loss, global_step=global_step); with tf.control_dependencies([train_step, variable_averages_op]): train_op = tf.no_op(name='train') # 初始化Tensorflow持久化类 saver = tf.train.Saver() with tf.Session() as sess: tf.initialize_all_variables().run() # 在训练过程中不再测试模型在验证数据上的表现，验证和测试的过程将会有一个独立的程序来完成 for i in range(TRAINING_STEPS): xs, ys = mnist.train.next_batch(BATCH_SIZE) _, loss_value, step = sess.run([train_op, loss, global_step], feed_dict={x:xs, y_:ys}) # 每1000轮保存一次模型 if i % 1000 == 0: # 输出当前训练情况。这里只输出了模型在当前训练batch上的损失函数大小 # 通过损失函数的大小可以大概了解训练的情况。在验证数据集上的正确率信息 # 会有一个单独的程序来生成 print("After %d training step(s),loss on training batch is %g" % (step, loss_value)) # 保存当前的模型。注意这里给出了global_step参数，这样可以让每个被保存模型的文件末尾加上训练的轮数 # 比如"model.ckpt-1000"表示训练1000轮之后得到的模型 saver.save(sess, os.path.join(MODEL_SAVE_PATH, MODEL_NAME), global_step=global_step) def main(argv=None): mnist = input_data.read_data_sets("/tmp/data", one_hot=True) train(mnist) if __name__ == '__main__': tf.app.run()
上面的代码是mnist_train.py里边
1、封装了网络滑动平均值的计算
2、网络误差的计算：信息的交叉熵+二范式的误差
3、误差的迭代优化算法，梯度下降
4、使用的激活函数是logist函数
5、对输出的结果进行了softmax处理，用来输出结果和真实值的交叉熵
6、持久化过程：将每隔1000次训练的结果保存到/path/to/model/文件夹下
7、然后在测试的时候读取这个结果graph，对测试数据进行测试
训练的结果如下所示
Extracting /tmp/data rain-images-idx3-ubyte.gz Extracting /tmp/data rain-labels-idx1-ubyte.gz Extracting /tmp/data 10k-images-idx3-ubyte.gz Extracting /tmp/data 10k-labels-idx1-ubyte.gz E c: f_jenkinshomeworkspace elease-windevicecpuoswindows ensorflowcoreframeworkop_kernel.cc:943] OpKernel ('op: "BestSplits" device_type: "CPU"') for unknown op: BestSplits E c: f_jenkinshomeworkspace elease-windevicecpuoswindows ensorflowcoreframeworkop_kernel.cc:943] OpKernel ('op: "CountExtremelyRandomStats" device_type: "CPU"') for unknown op: CountExtremelyRandomStats E c: f_jenkinshomeworkspace elease-windevicecpuoswindows ensorflowcoreframeworkop_kernel.cc:943] OpKernel ('op: "FinishedNodes" device_type: "CPU"') for unknown op: FinishedNodes E c: f_jenkinshomeworkspace elease-windevicecpuoswindows ensorflowcoreframeworkop_kernel.cc:943] OpKernel ('op: "GrowTree" device_type: "CPU"') for unknown op: GrowTree E c: f_jenkinshomeworkspace elease-windevicecpuoswindows ensorflowcoreframeworkop_kernel.cc:943] OpKernel ('op: "ReinterpretStringToFloat" device_type: "CPU"') for unknown op: ReinterpretStringToFloat E c: f_jenkinshomeworkspace elease-windevicecpuoswindows ensorflowcoreframeworkop_kernel.cc:943] OpKernel ('op: "SampleInputs" device_type: "CPU"') for unknown op: SampleInputs E c: f_jenkinshomeworkspace elease-windevicecpuoswindows ensorflowcoreframeworkop_kernel.cc:943] OpKernel ('op: "ScatterAddNdim" device_type: "CPU"') for unknown op: ScatterAddNdim E c: f_jenkinshomeworkspace elease-windevicecpuoswindows ensorflowcoreframeworkop_kernel.cc:943] OpKernel ('op: "TopNInsert" device_type: "CPU"') for unknown op: TopNInsert E c: f_jenkinshomeworkspace elease-windevicecpuoswindows ensorflowcoreframeworkop_kernel.cc:943] OpKernel ('op: "TopNRemove" device_type: "CPU"') for unknown op: TopNRemove E c: f_jenkinshomeworkspace elease-windevicecpuoswindows ensorflowcoreframeworkop_kernel.cc:943] OpKernel ('op: "TreePredictions" device_type: "CPU"') for unknown op: TreePredictions E c: f_jenkinshomeworkspace elease-windevicecpuoswindows ensorflowcoreframeworkop_kernel.cc:943] OpKernel ('op: "UpdateFertileSlots" device_type: "CPU"') for unknown op: UpdateFertileSlots WARNING:tensorflow:From C:UsersweizhenworkspaceTextUtilmnist_train.py:49: initialize_all_variables (from tensorflow.python.ops.variables) is deprecated and will be removed after 2017-03-02. Instructions for updating: Use `tf.global_variables_initializer` instead. After 1 training step(s),loss on training batch is 3.18871 After 1001 training step(s),loss on training batch is 0.236992 After 2001 training step(s),loss on training batch is 0.184191 After 3001 training step(s),loss on training batch is 0.141068 After 4001 training step(s),loss on training batch is 0.120774 After 5001 training step(s),loss on training batch is 0.11057 After 6001 training step(s),loss on training batch is 0.0951077 After 7001 training step(s),loss on training batch is 0.128405 After 8001 training step(s),loss on training batch is 0.0825269 After 9001 training step(s),loss on training batch is 0.0735326 After 10001 training step(s),loss on training batch is 0.0714079 After 11001 training step(s),loss on training batch is 0.0640013 After 12001 training step(s),loss on training batch is 0.0633993 After 13001 training step(s),loss on training batch is 0.0620059 After 14001 training step(s),loss on training batch is 0.0566063 After 15001 training step(s),loss on training batch is 0.0471778 After 16001 training step(s),loss on training batch is 0.0527759 After 17001 training step(s),loss on training batch is 0.0543462 After 18001 training step(s),loss on training batch is 0.0416694 After 19001 training step(s),loss on training batch is 0.0472751 After 20001 training step(s),loss on training batch is 0.0419436 After 21001 training step(s),loss on training batch is 0.0384671 After 22001 training step(s),loss on training batch is 0.0391581 After 23001 training step(s),loss on training batch is 0.0355204 After 24001 training step(s),loss on training batch is 0.0357683 After 25001 training step(s),loss on training batch is 0.0357087 After 26001 training step(s),loss on training batch is 0.0377925 After 27001 training step(s),loss on training batch is 0.0336796 After 28001 training step(s),loss on training batch is 0.0417044 After 29001 training step(s),loss on training batch is 0.0374377
对训练的模型进行测试：
''' Created on 2017年4月22日 @author: weizhen ''' import time import tensorflow as tf from tensorflow.examples.tutorials.mnist import input_data #加载mnist_inference.py和mnist_train.py中定义的常量和函数 import mnist_inference import mnist_train #每10秒加载一次最新的模型，并在测试数据上测试最新模型的正确率 EVAL_INTERVAL_SECS=10 def evaluate(mnist): with tf.Graph().as_default() as g: #定义输入输出的格式 x = tf.placeholder(tf.float32, [None,mnist_inference.INPUT_NODE], name='x-input') y_= tf.placeholder(tf.float32, [None,mnist_inference.OUTPUT_NODE], name='y-input') validate_feed={x:mnist.validation.images,y_:mnist.validation.labels} #直接通过调用封装好的函数来计算前向传播结果。因为测试时不关注正则化损失的值 #所以这里用于计算正则化损失的函数被设置为None y=mnist_inference.inference(x, None) #使用前向传播的结果计算正确率。如果需要对未知的样例进行分析，那么使用 #tf.argmax(y,1)就可以得到输入样例的预测类别了 correct_prediction=tf.equal(tf.argmax(y, 1), tf.argmax(y_, 1)) accuracy=tf.reduce_mean(tf.cast(correct_prediction, tf.float32)) #通过变量重命名的方式来加载模型，这样在前向传播的过程中就不需要调用求滑动平均的函数来获取平均值了 #这样就可以完全共用mnist_inference.py中定义的前向传播过程 variable_averages=tf.train.ExponentialMovingAverage(mnist_train.MOVING_AVERAGE_DECAY) variables_to_restore=variable_averages.variables_to_restore() saver = tf.train.Saver(variables_to_restore) #每隔EVAL_INTERVAL_SECS秒调用一次计算正确率的过程以检测训练过程中正确率的变化 while True: with tf.Session() as sess: #tf.train.get_checkpoint_state 函数会通过checkpoint文件自动找到目录中最新模型的文件名 ckpt=tf.train.get_checkpoint_state(mnist_train.MODEL_SAVE_PATH) if ckpt and ckpt.model_checkpoint_path: #加载模型 saver.restore(sess, ckpt.model_checkpoint_path) #通过文件名得到模型保存时迭代的轮数 global_step=ckpt.model_checkpoint_path.split('/')[-1].split('-')[-1] accuracy_score=sess.run(accuracy,feed_dict=validate_feed) print("After %s training step(s) validation accuracy=%g"%(global_step,accuracy_score)) else: print('No checkpoint file found') return time.sleep(EVAL_INTERVAL_SECS) def main(argv=None): mnist=input_data.read_data_sets("/tmp/data",one_hot=True) evaluate(mnist) if __name__=='__main__': tf.app.run()