TF:基于CNN(2+1)实现MNIST手写数字图片识别准确率提高到99%

输出结果


Extracting MNIST_data/train-images-idx3-ubyte.gz

Extracting MNIST_data/train-labels-idx1-ubyte.gz

Extracting MNIST_data/t10k-images-idx3-ubyte.gz

Extracting MNIST_data/t10k-labels-idx1-ubyte.gz

step 0, training accuracy 0.1

step 1000, training accuracy 0.98

step 2000, training accuracy 0.96

step 3000, training accuracy 1

step 4000, training accuracy 1

step 5000, training accuracy 0.98

step 6000, training accuracy 0.98

step 7000, training accuracy 1

step 8000, training accuracy 1

step 9000, training accuracy 1

step 10000, training accuracy 1

step 11000, training accuracy 1

step 12000, training accuracy 1

step 13000, training accuracy 0.98

step 14000, training accuracy 1

step 15000, training accuracy 1

step 16000, training accuracy 1

step 17000, training accuracy 1

step 18000, training accuracy 1

step 19000, training accuracy 1




代码实现


#TF:基于CNN实现MNIST手写数字识别准确率提高到99%

import tensorflow as tf

from tensorflow.examples.tutorials.mnist import input_data

……

if __name__ == '__main__':

   mnist = input_data.read_data_sets("MNIST_data/", one_hot=True)

   x = tf.placeholder(tf.float32, [None, 784])

   y_ = tf.placeholder(tf.float32, [None, 10])

   x_image = tf.reshape(x, [-1, 28, 28, 1])  #x_image就是输入的训练图像

   W_conv1 = weight_variable([5, 5, 1, 32])

   b_conv1 = bias_variable([32])

   h_conv1 = tf.nn.relu(conv2d(x_image, W_conv1) + b_conv1)  #是真正进行卷积计算,再选用ReLU作为激活函数

   h_pool1 = max_pool_2x2(h_conv1)  #调用函数max_pool_2x2 进行一次池化操作。

 

   W_conv2 = weight_variable([5, 5, 32, 64])

   b_conv2 = bias_variable([64])

   h_conv2 = tf.nn.relu(conv2d(h_pool1, W_conv2) + b_conv2)

   h_pool2 = max_pool_2x2(h_conv2)

   W_fc1 = weight_variable([7 * 7 * 64, 1024])

   b_fc1 = bias_variable([1024])

   h_pool2_flat = tf.reshape(h_pool2, [-1, 7 * 7 * 64])

   h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)

   keep_prob = tf.placeholder(tf.float32)

   h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)

   W_fc2 = weight_variable([1024, 10])

   b_fc2 = bias_variable([10])

   y_conv = tf.matmul(h_fc1_drop, W_fc2) + b_fc2  

   cross_entropy = tf.reduce_mean(

       tf.nn.softmax_cross_entropy_with_logits(labels=y_, logits=y_conv))

   train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy)

   correct_prediction = tf.equal(tf.argmax(y_conv, 1), tf.argmax(y_, 1))

   accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))

   sess = tf.InteractiveSession()

   sess.run(tf.global_variables_initializer())

   for i in range(20000):  # 训练20000步

       batch = mnist.train.next_batch(50)

       # 每100步报告一次在验证集上的准确度

       if i % 100 == 0:

           train_accuracy = accuracy.eval(feed_dict={

               x: batch[0], y_: batch[1], keep_prob: 1.0})

           print("step %d, training accuracy %g" % (i, train_accuracy))

       train_step.run(feed_dict={x: batch[0], y_: batch[1], keep_prob: 0.5})

   print("test accuracy %g" % accuracy.eval(feed_dict={

       x: mnist.test.images, y_: mnist.test.labels, keep_prob: 1.0}))


上一篇:Web项目开发实践,Cookie与Session机制


下一篇:TF:利用是Softmax回归+GD算法实现MNIST手写数字图片识别(10000张图片测试得到的准确率为92%)