卷积神经网络python代码

# Lab 11 MNIST and Convolutional Neural Network import tensorflow as tf import random import scipy.io as scio # import matplotlib.pyplot as plt from tensorflow.examples.tutorials.mnist import input_data tf.set_random_seed(777) # reproducibility mnist = input_data.read_data_sets("MNIST_data/", one_hot=True) # Check out https://www.tensorflow.org/get_started/mnist/beginners for # more information about the mnist dataset # hyper parameters learning_rate = 0.002 training_epochs = 3 batch_size = 50 # input place holders X = tf.placeholder(tf.float32, [None, 784]) X_img = tf.reshape(X, [-1, 28, 28, 1]) # img 28x28x1 (black/white) Y = tf.placeholder(tf.float32, [None, 10]) # L1 ImgIn shape=(?, 28, 28, 1) W1 = tf.Variable(tf.random_normal([5, 5, 1, 6], stddev=0.01)) b1 = tf.Variable(tf.random_normal([6])) # Conv -> (?, 24, 24, 6) # Pool -> (?, 12, 12, 6) L1 = tf.nn.conv2d(X_img, W1, strides=[1, 1, 1, 1], padding='VALID') L1 = tf.nn.relu(L1 + b1) L1 = tf.nn.avg_pool(L1, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME') ''' Tensor("Conv2D:0", shape=(?, 24, 24, 6), dtype=float32) Tensor("Relu:0", shape=(?, 24, 24, 6), dtype=float32) Tensor("avgPool:0", shape=(?, 12, 12, 6), dtype=float32) ''' # L2 ImgIn shape=(?, 12, 12, 6) W2 = tf.Variable(tf.random_normal([5, 5, 6, 16], stddev=0.01)) b2 = tf.Variable(tf.random_normal([16])) # Conv ->(?, 8, 8, 16) # Pool ->(?, 4, 4, 16) L2 = tf.nn.conv2d(L1, W2, strides=[1, 1, 1, 1], padding='VALID') L2 = tf.nn.relu(L2 + b2) L2 = tf.nn.avg_pool(L2, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME') L2_flat = tf.reshape(L2, [-1, 4 * 4 * 16]) #L2_flat = tf.nn.relu(L2_flat) ''' Tensor("Conv2D_1:0", shape=(?, 14, 14, 64), dtype=float32) Tensor("Relu_1:0", shape=(?, 14, 14, 64), dtype=float32) Tensor("MaxPool_1:0", shape=(?, 7, 7, 64), dtype=float32) Tensor("Reshape_1:0", shape=(?, 3136), dtype=float32) ''' # Final FC 7x7x64 inputs -> 10 outputs W3 = tf.Variable(tf.random_normal([4*4*16, 10], stddev=0.01)) #W3 = tf.get_variable("W3", shape=[4 * 4 * 16, 10], # initializer=tf.contrib.layers.xavier_initializer()) b3 = tf.Variable(tf.random_normal([10])) logits = tf.matmul(L2_flat, W3) + b3 #logits = tf.nn.relu(logits) # define cost/loss & optimizer cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits( logits=logits, labels=Y)) optimizer = tf.train.AdamOptimizer(learning_rate=learning_rate).minimize(cost) #creat a saver saver = tf.train.Saver() # initialize with tf.Session() as sess: sess.run(tf.global_variables_initializer()) # train my model print('Learning started. It takes sometime.') for epoch in range(training_epochs): avg_cost = 0 total_batch = int(mnist.train.num_examples / batch_size) for i in range(total_batch): batch_xs, batch_ys = mnist.train.next_batch(batch_size) feed_dict = {X: batch_xs, Y: batch_ys} c, _ = sess.run([cost, optimizer], feed_dict=feed_dict) avg_cost += c / total_batch print('Epoch:', '%04d' % (epoch + 1), 'cost =', '{:.9f}'.format(avg_cost)) print('Learning Finished!') # save the model saver.save(sess, "model/save_net.ckpt") # save_mat # dataW1 = 'w1.mat' # scio.savemat(dataW1, {'W1':sess.run(W1)}) # dataW2 = 'w2.mat' # scio.savemat(dataW2, {'W2':sess.run(W2)}) # dataW3 = 'w3.mat' # scio.savemat(dataW3, {'W3':sess.run(W3)}) # datab1 = 'b1.mat' # scio.savemat(datab1, {'b1':sess.run(b1)}) # datab2 = 'b2.mat' # scio.savemat(datab2, {'b2':sess.run(b2)}) # datab3 = 'b3.mat' # scio.savemat(datab3, {'b3':sess.run(b3)}) # Test model and check accuracy correct_prediction = tf.equal(tf.argmax(logits, 1), tf.argmax(Y, 1)) accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32)) print('Accuracy:', sess.run(accuracy, feed_dict={ X: mnist.test.images, Y: mnist.test.labels})) # Get one and predict r = random.randint(0, mnist.test.num_examples - 1) print("Label: ", sess.run(tf.argmax(mnist.test.labels[r:r + 1], 1))) print("Prediction: ", sess.run( tf.argmax(logits, 1), feed_dict={X: mnist.test.images[r:r + 1]})) # plt.imshow(mnist.test.images[r:r + 1]. # reshape(28, 28), cmap='Greys', interpolation='nearest') # plt.show() ''' Epoch: 0001 cost = 0.340291267 Epoch: 0002 cost = 0.090731326 Epoch: 0003 cost = 0.064477619 Epoch: 0004 cost = 0.050683064 Epoch: 0005 cost = 0.041864835 Epoch: 0006 cost = 0.035760704 Epoch: 0007 cost = 0.030572132 Epoch: 0008 cost = 0.026207981 Epoch: 0009 cost = 0.022622454 Epoch: 0010 cost = 0.019055919 Epoch: 0011 cost = 0.017758641 Epoch: 0012 cost = 0.014156652 Epoch: 0013 cost = 0.012397016 Epoch: 0014 cost = 0.010693789 Epoch: 0015 cost = 0.009469977 Learning Finished! Accuracy: 0.9885 ''' ''' Epoch: 0001 cost = 0.318063878 Epoch: 0002 cost = 0.105674432 Epoch: 0003 cost = 0.077878999 Epoch: 0004 cost = 0.066057363 Epoch: 0005 cost = 0.057512918 Epoch: 0006 cost = 0.052459399 Epoch: 0007 cost = 0.046069056 Epoch: 0008 cost = 0.042238329 Epoch: 0009 cost = 0.039322835 Epoch: 0010 cost = 0.034931831 Epoch: 0011 cost = 0.033192076 Epoch: 0012 cost = 0.030914000 Epoch: 0013 cost = 0.028490152 Epoch: 0014 cost = 0.026866245 Epoch: 0015 cost = 0.025608637 Epoch: 0016 cost = 0.024366001 Epoch: 0017 cost = 0.021244116 Epoch: 0018 cost = 0.022134637 Epoch: 0019 cost = 0.020379513 Epoch: 0020 cost = 0.018623088 Epoch: 0021 cost = 0.018348278 Epoch: 0022 cost = 0.017074608 Epoch: 0023 cost = 0.016454066 Epoch: 0024 cost = 0.014578759 Epoch: 0025 cost = 0.016178100 Epoch: 0026 cost = 0.013665723 Epoch: 0027 cost = 0.013868410 Epoch: 0028 cost = 0.013506715 Epoch: 0029 cost = 0.012520027 Epoch: 0030 cost = 0.011752283 Epoch: 0031 cost = 0.012224808 Epoch: 0032 cost = 0.009687688 Epoch: 0033 cost = 0.011666438 Epoch: 0034 cost = 0.010130068 Epoch: 0035 cost = 0.009605887 Epoch: 0036 cost = 0.009520990 Epoch: 0037 cost = 0.009449142 Epoch: 0038 cost = 0.009283410 Epoch: 0039 cost = 0.007676879 Epoch: 0040 cost = 0.008369593 Epoch: 0041 cost = 0.009352039 Epoch: 0042 cost = 0.007622884 Epoch: 0043 cost = 0.008005746 Epoch: 0044 cost = 0.006593210 Epoch: 0045 cost = 0.006523430 Epoch: 0046 cost = 0.007701061 Epoch: 0047 cost = 0.005825046 Epoch: 0048 cost = 0.007374606 Epoch: 0049 cost = 0.006955691 Epoch: 0050 cost = 0.005048158 Epoch: 0051 cost = 0.006731462 Epoch: 0052 cost = 0.006690387 Epoch: 0053 cost = 0.004819284 Epoch: 0054 cost = 0.005591960 Epoch: 0055 cost = 0.005885528 Epoch: 0056 cost = 0.005811614 Epoch: 0057 cost = 0.005956684 Epoch: 0058 cost = 0.004923902 Epoch: 0059 cost = 0.005496770 Epoch: 0060 cost = 0.005378819 Learning Finished! Accuracy: 0.99 '''