手写识别

Tensorflow_手写识别.py

from tensorflow.examples.tutorials.mnist import input_data
from PIL import Image
import numpy as np
#路径要写全
MNIST_data_folder="D:\MNIST\mnist"
mnist=input_data.read_data_sets(MNIST_data_folder,one_hot=True)
print(mnist.train.images.shape, mnist.train.labels.shape)
print(mnist.test.images.shape, mnist.test.labels.shape)
print(mnist.validation.images.shape, mnist.validation.labels.shape)

import tensorflow as tf

# 定义两个placeholder
x = tf.placeholder(tf.float32, [None, 784])
y = tf.placeholder(tf.float32, [None, 10])
x1 = tf.placeholder(tf.float32, [None, 784])

keep_prob = tf.placeholder("float")

# 创建一个简单的神经网络
W = tf.Variable(tf.zeros([784, 10]))
b = tf.Variable(tf.zeros([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)

# 结果存放在一个布尔型列表中
correct_prediction = tf.equal(tf.argmax(y, 1), tf.argmax(prediction, 1))  # argmax返回一维张量中最大的值所在的位置
# 求准确率
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))

# 创建 summary node
# w, b 画直方图
# loss, accuracy画标量图
tf.summary.histogram('weight', W)
tf.summary.histogram('bias', b)
tf.summary.scalar('loss', loss)
tf.summary.scalar('accuracy', accuracy)
merged_summary_op = tf.summary.merge_all()

# 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
print(n_batch)
logs_path = 'C:\\Users\\R\\Desktop\\'

saver = tf.train.Saver() #定义saver

CKPT_DIR = 'C:\\Users\\R\\Desktop\\'
ckpt = tf.train.get_checkpoint_state(CKPT_DIR)





with tf.Session() as sess:


    init = tf.global_variables_initializer()
    sess.run(init)
   # saver.restore(sess, "d:/MNISTSAVE/model.ckpt")

    ckpt = tf.train.get_checkpoint_state(CKPT_DIR)
    if ckpt and ckpt.model_checkpoint_path:
     saver.restore(sess, ckpt.model_checkpoint_path)



    summary_writer = tf.summary.FileWriter(logs_path, graph=tf.get_default_graph())
    for epoch in range(0):
        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))
        if epoch % 100 == 0:
            saver.save(sess, 'C:\\Users\\R\\Desktop\\' + 'model', global_step=epoch)

       # saver.save(sess, 'd:/MNISTSAVE/model.ckpt')  # 模型储存位置
        # summary_writer.add_summary(acc, epoch * n_batch + batch)
    image_path="C:\\Users\\R\\Desktop\\3_59.png"
    img = Image.open(image_path).convert('L')
    flatten_img = np.reshape(img, 784)
    x1 = np.array([1 - flatten_img])
    output = sess.run(prediction, feed_dict={x: x1, keep_prob:1.0})
   # out = sess.run(y, feed_dict={x:x1})
    print(image_path)
    print('        -> Predict digit', np.argmax(output))