TensorFlow uses CNN to analyze mnist handwritten digital data sets

Import TensorFlow as TF import numpy as NP import OS os.environ[' tf_cpp_min_log_level '] = ' 2 ' from Tensorflow.examples.tut Orials.mnist import Input_data mnist = Input_data.read_data_sets ("mnist_data/", One_hot=true) TrX, TrY, TeX, TeY = mnist. Train.images, Mnist.train.labels, Mnist.test.images, Mnist.test.labels #把上述trX和teX的形状变为 [ -1,28,28,1],-1 indicates that the number of input pictures is not considered
, 28x28 is a picture of the long and wide pixels, # 1 is the number of channels (channel), because the mnist picture is black and white, so the channel is 1, if the RGB color image, the channel is 3. TrX = Trx.reshape ( -1, 1) # 28x28x1 input img teX = Tex.reshape ( -1,, 1) # 28x28x1 input img X = TF.PLACEH
Older ("float", [None, 1]) Y = Tf.placeholder ("float", [None,]) #初始化权重与定义网络结构. # Here, we're going to build a convolution neural network def init_weights (shape) with 3 convolution layers and 3 pool layers, followed by 1 full join layers and one output layer: return TF.  Variable (Tf.random_normal (Shape, stddev=0.01)) W = init_weights ([3, 3, 1, M]) # patch size is 3x3, input dimension is 1, output dimension is W2 = Init_weights ([3, 3,,]) # patch size is 3x3, input dimension is 32, output dimension is W3 = Init_weights ([3, 3,, 128]) # Patch Big Small for 3x3, input dimension of 64, output dimension is 128
W4 = Init_weights ([128 * 4 * 4, 625]) # Fully connected layer, input dimension is 128x4x4, the output data of the upper layer is transformed into one-dimensional, the output dimension is 625 w_o = init_weights ([625, 10]) # output layer, the input dimension is 625, the output dimension is 10, representing the 10 Class (labels) # Neural network model constructor, passing in the following parameters # X: Input data # W: weight of each layer # P_keep_conv,p_keep_hidden:dropout To preserve the neuron ratio def model (X, W, W2, W3, W4, W_o, P_keep_conv, P_keep_hidden): # The first set of convolution layer and pool layer, and finally dropout some neurons l1a = TF.NN.R  Elu (tf.nn.conv2d (X, W, strides=[1, 1, 1, 1], padding= ' SAME ')) # l1a shape= (?, M, m) L1 = Tf.nn.max_pool (l1a, Ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding= ' SAME ') # L1 shape= (?,,) L1 = tf.nn.dropout (L1, P_kee
    P_CONV) # Second set of convolution layer and pool layer, finally dropout some neurons l2a = Tf.nn.relu (tf.nn.conv2d (L1, W2, Strides=[1, 1, 1, 1], padding= ' SAME '))
    # l2a shape=, L2 = Tf.nn.max_pool (l2a, Ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding= ' SAME ') # L2 Shape= (?, 7, 7,) L2 = Tf.nn.dropout (L2, P_keep_conv) # The third set of convolution layer and pool layer, and finally dropout some neurons l3a = Tf.nn.relu (TF. nn.conv2d (L2, W3, StrideS=[1, 1, 1, 1], padding= ' SAME ') # l3a shape= (?, 7, 7, 128) L3 = Tf.nn.max_pool (l3a, Ksize=[1, 2, 2, 1], strides=[  1, 2, 2, 1], padding= ' SAME ') # L3 shape= (?, 4, 4, 128) L3 = Tf.reshape (L3, [-1, W4.get_shape (). As_list () [0]]) Reshape to (?, 2048) L3 = Tf.nn.dropout (L3, P_keep_conv) # fully connected layer, and finally dropout some neurons L4 = Tf.nn.relu (Tf.matmul (L3, W4) L4 = Tf.nn.dropout (L4, P_keep_hidden) # output Layer Pyx = Tf.matmul (L4, W_o) return Pyx #返回预测值 #我们定义dropout Placeholder--keep_conv, which indicates how much of the neuron is preserved in the first layer. Generate network model, get predictive value P_keep_conv = Tf.placeholder ("float") P_keep_hidden = Tf.placeholder ("float") py_x = Model (x, W, W2, W3, W4, W_o, P_keep_conv, P_keep_hidden) #得到预测值 #定义损失函数, here we still use the tf.nn.softmax_cross_entropy_with_logits to compare the difference between predicted and real values, and do mean processing; # define the operation of the Training (TRAIN_OP), using the Rmsprop algorithm to achieve the optimizer Tf.train.RMSPropOptimizer, the learning rate is 0.001, the attenuation value of 0.9, so that the loss of the smallest; # defines the operation of the forecast (predict_ OP) cost = Tf.reduce_mean (Tf.nn softmax_cross_entropy_with_logits (logits=py_x, labels=y)) Train_op = Tf.train.RMSPropOptimizer (0001, 0.9). Minimize (cost) Predict_op = Tf.argmax (py_x, 1) #定义训练时的批次大小和评估时的批次大小 batch_size = 128 test_size = 256 #在一个会话中启动图 , begin training and evaluation # Launch the graph in a session with TF. Session () as Sess: # Your need to initialize all variables TF. 
                             Global_variables_initializer (). Run () for I in Range (m): Training_batch = Zip (range (0, Len (trX), batch_size),
            Range (Batch_size, Len (TrX) +1, batch_size)) for start, end in Training_batch: Sess.run (Train_op, Feed_dict={x:trx[start:end], y:try[start:end], P_keep_ conv:0.8, p_keep_hidden:0.5}) test_indices = Np.arange (len (TeX)) # get A test Batch np.random.shuffle ( test_indices) test_indices = test_indices[0:test_size] Print (i, Np.mean (Np.argmax (tey[test_indices), axis
                                                         =1) = = Sess.run (Predict_op, feed_dict={x:tex[test_indices], P_keep_conv:1.0, p_keep_hidden:1.0}))