本章内容:
1.tf.nn.conv2d()--卷积函数的运算过程
2.tensorflow--线性回归
3.tensorflow--逻辑回归
4.浅层神经网络(sgimoid)
5.浅层神经网络+一个隐藏层(relu函数)
6.浅层神经网络+cnn卷积)
7.训练卷积神经网络+保存模型
8.训练卷积神经网络+恢复模型
1.验证卷积函数的运算过程
"""
函数格式:tf.nn.conv2d(input, filter, strides, padding, use_cudnn_on_gpu = Noen, name = None)
1.input:指需要做卷积的输入图像
要求是一个4维的Tensor,shape为[batch, in_height, in_width, in_channels]:
batch:训练时一个batch的图片数量
in_height:输入图像的高度
in_width:输入图像的宽度
in_channels:输入图像的通道数,灰度图像则为1,彩色图像则为3
2.filter:CNN卷积网络中的卷积核
要求是一个Tensor,类型和input类型相同,shape为[filter_height, filter_width, in_channels, out_channels]:
filter_height:卷积核的高度
filter_width:卷积核的宽度
in_channels:图像的通道数,input的in_channels相同
out_channels:卷积核的个数
3.strides:不同维度上的步长,是一个一维向量,[ 1, strides, strides, 1],
第一维和最后一维的数字要求必须是1。因为卷积层的步长只对矩阵的长和宽有效。
4.padding:string类型,表示卷积的形式,是否考虑边界,值为“SAME”和“VALID”
"SAME"是考虑边界,不足的时候用填充周围(即input为多大,output就多大),"VALID"则不考虑边界。
5.use_cudnn_on_gpu: bool类型,是否使用cudnn加速,默认为true。
"""
# #case1:简单卷积实验
# #输入是1张 3*3 大小的图片,图像通道数是5,卷积核是 1*1 大小,数量是1 ,步长是[1,1,1,1],
# #最后得到一个 3*3 的feature map。1张图最后输出就是一个 shape为[1,3,3,1] 的tensor。
#
# # -*- coding:utf-8 -*-
# import tensorflow as tf
# Input = tf.Variable(tf.random_normal([1, 3, 3, 5]))
# Filter = tf.Variable(tf.random_normal([1, 1, 5, 1]))
# conv1 = tf.nn.conv2d(Input, Filter, strides=[1, 1, 1, 1], padding='VALID')
# with tf.Session() as sess:
# # 初始化变量
# op_init = tf.global_variables_initializer()
# sess.run(op_init)
# a=sess.run(conv1)
# print(a.shape)
# #case2:"VALID"、"SAME"的区别
# # 输入是1张 3*3 大小的图片,图像通道数是5,卷积核是 2*2大小,数量是1 ,步长是[1,1,1,1],
# # padding 设置为“VALID”,最后得到一个 2*2的feature map;“SAME”得到3*3的feature
# #
# # -*- coding:utf-8 -*-
# import tensorflow as tf
# Input = tf.Variable(tf.random_normal([1, 3, 3, 5]))
# Filter = tf.Variable(tf.random_normal([2, 2, 5, 1]))
# conv2 = tf.nn.conv2d(Input, Filter, strides=[1, 1, 1, 1], padding='VALID') #不自动填充边界
# conv3 = tf.nn.conv2d(Input, Filter, strides=[1, 1, 1, 1], padding='SAME') #自动填充边界
# with tf.Session() as sess:
# # 初始化变量
# op_init = tf.global_variables_initializer()
# sess.run(op_init)
# print("不自动填充边界:",sess.run(conv2).shape)
# print("自动填充边界:",sess.run(conv3).shape)
# #case3:输入多张图片的情况
# # 输入是2张 3*3 大小的图片,图像通道数是5,卷积核是 2*2大小,数量是1,步长是[1,1,1,1],
# # padding 设置为“SAME”,最后得到2个 3*3的feature map。1张图最后输出就是一个 shape为[2,3,3,1] 的tensor。
# # -*- coding:utf-8 -*-
# import tensorflow as tf
# Input = tf.Variable(tf.random_normal([2, 3, 3, 5]))
# Filter = tf.Variable(tf.random_normal([2, 2, 5, 1]))
# conv4 = tf.nn.conv2d(Input, Filter, strides=[1, 1, 1, 1], padding='SAME')
# with tf.Session() as sess:
# # 初始化变量
# op_init = tf.global_variables_initializer()
# sess.run(op_init)
# print(sess.run(conv4).shape)
# case4:输入多张图片、多个卷积核的情况
# 输入是4张 3*3 大小的图片,图像通道数是5,卷积核是 2*2大小,数量是4,步长是[1,1,1,1],
# padding 设置为“SAME”,最后每张图片得到4个 3*3的feature map。最后输出是shape为[4,3,3,4] 的tensor。
# -*- coding:utf-8 -*-
import tensorflow as tf
Input = tf.Variable(tf.random_normal([4, 3, 3, 5]))
Filter = tf.Variable(tf.random_normal([2, 2, 5, 4]))
conv5 = tf.nn.conv2d(Input, Filter, strides=[1, 1, 1, 1], padding='SAME')
with tf.Session() as sess:
# 初始化变量
op_init = tf.global_variables_initializer()
sess.run(op_init)
print(sess.run(conv5).shape)
2.tensorflow实现线性回归
import tensorflow as tf
import numpy as np
trX = np.linspace(-1, 1, 101)
trY = 2 * trX + np.random.randn(*trX.shape) * 0.33 # 创建一些线性值附近的随机值
X = tf.placeholder("float")
Y = tf.placeholder("float")
def model(X, w):
return tf.multiply(X, w) # X*w线性求值,非常简单
w = tf.Variable(0.0, name="weights")
y_model = model(X, w)
cost = tf.square(Y - y_model) # 用平方误差做为优化目标
train_op = tf.train.GradientDescentOptimizer(0.01).minimize(cost) # 梯度下降优化
# 开始创建Session干活!
with tf.Session() as sess:
# 首先需要初始化全局变量,这是Tensorflow的要求
tf.global_variables_initializer().run()
for i in range(100):
for (x, y) in zip(trX, trY):
sess.run(train_op, feed_dict={X: x, Y: y})
print(sess.run(w))
3.tensorflow实现逻辑回归
import tensorflow as tf
import numpy as np
from tensorflow.examples.tutorials.mnist import input_data
def init_weights(shape):
return tf.Variable(tf.random_normal(shape, stddev=0.01))
def model(X, w):
return tf.matmul(X, w) # 模型还是矩阵乘法
#加载mnist数据集
mnist = input_data.read_data_sets("mnist/", one_hot=True)
trX, trY, teX, teY = mnist.train.images, mnist.train.labels, mnist.test.images, mnist.test.labels
X = tf.placeholder("float", [None, 784]) #784维,图片大小28*28
Y = tf.placeholder("float", [None, 10]) #10个标签,0-9
w = init_weights([784, 10])
py_x = model(X, w) #使用模型进行计算
cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=py_x, labels=Y)) # 计算误差
train_op = tf.train.GradientDescentOptimizer(0.05).minimize(cost) # construct optimizer
predict_op = tf.argmax(py_x, 1) #从结果py_x中找出每行最大的值的下标
with tf.Session() as sess:
tf.global_variables_initializer().run()
for i in range(10):
for start, end in zip(range(0, len(trX), 128), range(128, len(trX)+1, 128)):
sess.run(train_op, feed_dict={X: trX[start:end], Y: trY[start:end]})
print(i, "精确度:",np.mean(np.argmax(teY, axis=1) == sess.run(predict_op, feed_dict={X: teX})))
4.实现浅层神经网络(sgimoid)
import tensorflow as tf
import numpy as np
from tensorflow.examples.tutorials.mnist import input_data
# 所有连接随机生成权值
def init_weights(shape):
return tf.Variable(tf.random_normal(shape, stddev=0.01))
#跟之前比只增加了一个隐藏层:sigmoid函数,精度提高了
def model(X, w_h, w_o):
h = tf.nn.sigmoid(tf.matmul(X, w_h))
return tf.matmul(h, w_o)
mnist = input_data.read_data_sets("mnist/", one_hot=True)
trX, trY, teX, teY = mnist.train.images, mnist.train.labels, mnist.test.images, mnist.test.labels
X = tf.placeholder("float", [None, 784])
Y = tf.placeholder("float", [None, 10])
w_h = init_weights([784, 625])
w_o = init_weights([625, 10])
py_x = model(X, w_h, w_o)
cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=py_x, labels=Y)) # 计算误差损失
train_op = tf.train.GradientDescentOptimizer(0.05).minimize(cost) # construct an optimizer
predict_op = tf.argmax(py_x, 1)
with tf.Session() as sess:
tf.global_variables_initializer().run()
for i in range(100):
for start, end in zip(range(0, len(trX), 128), range(128, len(trX)+1, 128)):
sess.run(train_op, feed_dict={X: trX[start:end], Y: trY[start:end]})
print(i, np.mean(np.argmax(teY, axis=1) ==
sess.run(predict_op, feed_dict={X: teX})))
5.浅层神经网络+一个隐藏层(relu函数)
import tensorflow as tf
import numpy as np
from tensorflow.examples.tutorials.mnist import input_data
def init_weights(shape):
return tf.Variable(tf.random_normal(shape, stddev=0.01))
#增加了两个隐藏层:relu函数,精度提高
def model(X, w_h, w_h2, w_o, p_keep_input, p_keep_hidden):
X = tf.nn.dropout(X, p_keep_input)
h = tf.nn.relu(tf.matmul(X, w_h))
h = tf.nn.dropout(h, p_keep_hidden)
h2 = tf.nn.relu(tf.matmul(h, w_h2))
h2 = tf.nn.dropout(h2, p_keep_hidden)
return tf.matmul(h2, w_o)
mnist = input_data.read_data_sets("mnist/", one_hot=True)
trX, trY, teX, teY = mnist.train.images, mnist.train.labels, mnist.test.images, mnist.test.labels
X = tf.placeholder("float", [None, 784])
Y = tf.placeholder("float", [None, 10])
w_h = init_weights([784, 625])
w_h2 = init_weights([625, 625])
w_o = init_weights([625, 10])
p_keep_input = tf.placeholder("float")
p_keep_hidden = tf.placeholder("float")
py_x = model(X, w_h, w_h2, w_o, p_keep_input, p_keep_hidden)
cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=py_x, labels=Y))
train_op = tf.train.RMSPropOptimizer(0.001, 0.9).minimize(cost)
predict_op = tf.argmax(py_x, 1)
with tf.Session() as sess:
# you need to initialize all variables
tf.global_variables_initializer().run()
for i in range(100):
for start, end in zip(range(0, len(trX), 128), range(128, len(trX)+1, 128)):
sess.run(train_op, feed_dict={X: trX[start:end], Y: trY[start:end],
p_keep_input: 0.8, p_keep_hidden: 0.5})
print(i, np.mean(np.argmax(teY, axis=1) ==
sess.run(predict_op, feed_dict={X: teX,
p_keep_input: 1.0,
p_keep_hidden: 1.0})))
6.浅层神经网络+cnn卷积
import tensorflow as tf
import numpy as np
from tensorflow.examples.tutorials.mnist import input_data
batch_size = 128
test_size = 256
def init_weights(shape):
return tf.Variable(tf.random_normal(shape, stddev=0.01))
#调用了卷积神经网络:cnn,精度提高
def model(X, w, w2, w3, w4, w_o, p_keep_conv, p_keep_hidden):
l1a = tf.nn.relu(tf.nn.conv2d(X, w, # l1a shape=(?, 28, 28, 32)
strides=[1, 1, 1, 1], padding='SAME'))
l1 = tf.nn.max_pool(l1a, ksize=[1, 2, 2, 1], # l1 shape=(?, 14, 14, 32)
strides=[1, 2, 2, 1], padding='SAME')
l1 = tf.nn.dropout(l1, p_keep_conv)
l2a = tf.nn.relu(tf.nn.conv2d(l1, w2, # l2a shape=(?, 14, 14, 64)
strides=[1, 1, 1, 1], padding='SAME'))
l2 = tf.nn.max_pool(l2a, ksize=[1, 2, 2, 1], # l2 shape=(?, 7, 7, 64)
strides=[1, 2, 2, 1], padding='SAME')
l2 = tf.nn.dropout(l2, p_keep_conv)
l3a = tf.nn.relu(tf.nn.conv2d(l2, w3, # l3a shape=(?, 7, 7, 128)
strides=[1, 1, 1, 1], padding='SAME'))
l3 = tf.nn.max_pool(l3a, ksize=[1, 2, 2, 1], # l3 shape=(?, 4, 4, 128)
strides=[1, 2, 2, 1], padding='SAME')
l3 = tf.reshape(l3, [-1, w4.get_shape().as_list()[0]]) # reshape to (?, 2048)
l3 = tf.nn.dropout(l3, p_keep_conv)
l4 = tf.nn.relu(tf.matmul(l3, w4))
l4 = tf.nn.dropout(l4, p_keep_hidden)
pyx = tf.matmul(l4, w_o)
return pyx
mnist = input_data.read_data_sets("mnist/", one_hot=True)
trX, trY, teX, teY = mnist.train.images, mnist.train.labels, mnist.test.images, mnist.test.labels
trX = trX.reshape(-1, 28, 28, 1) # 28x28x1 input img
teX = teX.reshape(-1, 28, 28, 1) # 28x28x1 input img
X = tf.placeholder("float", [None, 28, 28, 1])
Y = tf.placeholder("float", [None, 10])
w = init_weights([3, 3, 1, 32]) # 3x3x1 conv, 32 outputs
w2 = init_weights([3, 3, 32, 64]) # 3x3x32 conv, 64 outputs
w3 = init_weights([3, 3, 64, 128]) # 3x3x32 conv, 128 outputs
w4 = init_weights([128 * 4 * 4, 625]) # FC 128 * 4 * 4 inputs, 625 outputs
w_o = init_weights([625, 10]) # FC 625 inputs, 10 outputs (labels)
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)
cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=py_x, labels=Y))
train_op = tf.train.RMSPropOptimizer(0.001, 0.9).minimize(cost)
predict_op = tf.argmax(py_x, 1)
with tf.Session() as sess:
# you need to initialize all variables
tf.global_variables_initializer().run()
for i in range(16):
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})))
7.训练卷积神经网络+保存模型
import tensorflow as tf
from tensorflow.examples.tutorials.mnist import input_data
import numpy as np
import os
# 屏蔽waring信息
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
"""------------------加载数据---------------------"""
# 载入数据
mnist = input_data.read_data_sets("mnist/", one_hot=True)
trX, trY, teX, teY = mnist.train.images, mnist.train.labels, mnist.test.images, mnist.test.labels
# 改变数据格式,为了能够输入卷积层
trX = trX.reshape(-1, 28, 28, 1) # -1表示不考虑输入图片的数量,1表示单通道
teX = teX.reshape(-1, 28, 28, 1)
"""------------------构建模型---------------------"""
# 定义输入输出的数据容器
X = tf.placeholder("float", [None, 28, 28, 1])
Y = tf.placeholder("float", [None, 10])
# 定义和初始化权重、dropout参数
def init_weights(shape):
return tf.Variable(tf.random_normal(shape, stddev=0.01))
w1 = init_weights([3, 3, 1, 32]) # 3X3的卷积核,获得32个特征
w2 = init_weights([3, 3, 32, 64]) # 3X3的卷积核,获得64个特征
w3 = init_weights([3, 3, 64, 128]) # 3X3的卷积核,获得128个特征
w4 = init_weights([128 * 4 * 4, 625]) # 从卷积层到全连层
w_o = init_weights([625, 10]) # 从全连层到输出层
p_keep_conv = tf.placeholder("float")
p_keep_hidden = tf.placeholder("float")
# 定义模型
def create_model(X, w1, w2, w3, w4, w_o, p_keep_conv, p_keep_hidden):
# 第一组卷积层和pooling层
conv1 = tf.nn.conv2d(X, w1, strides=[1, 1, 1, 1], padding='SAME')
conv1_out = tf.nn.relu(conv1)
pool1 = tf.nn.max_pool(conv1_out, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
pool1_out = tf.nn.dropout(pool1, p_keep_conv)
# 第二组卷积层和pooling层
conv2 = tf.nn.conv2d(pool1_out, w2, strides=[1, 1, 1, 1], padding='SAME')
conv2_out = tf.nn.relu(conv2)
pool2 = tf.nn.max_pool(conv2_out, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
pool2_out = tf.nn.dropout(pool2, p_keep_conv)
# 第三组卷积层和pooling层
conv3 = tf.nn.conv2d(pool2_out, w3, strides=[1, 1, 1, 1], padding='SAME')
conv3_out = tf.nn.relu(conv3)
pool3 = tf.nn.max_pool(conv3_out, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
pool3 = tf.reshape(pool3, [-1, w4.get_shape().as_list()[0]]) # 转化成一维的向量
pool3_out = tf.nn.dropout(pool3, p_keep_conv)
# 全连层
fully_layer = tf.matmul(pool3_out, w4)
fully_layer_out = tf.nn.relu(fully_layer)
fully_layer_out = tf.nn.dropout(fully_layer_out, p_keep_hidden)
# 输出层
out = tf.matmul(fully_layer_out, w_o)
return out
model = create_model(X, w1, w2, w3, w4, w_o, p_keep_conv, p_keep_hidden)
# 定义代价函数、训练方法、预测操作
cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=model, labels=Y))
train_op = tf.train.RMSPropOptimizer(0.001, 0.9).minimize(cost)
predict_op = tf.argmax(model, 1,name="predict")
# 定义一个saver
saver=tf.train.Saver()
# 定义存储路径
ckpt_dir="./ckpt_dir2"
if not os.path.exists(ckpt_dir):
os.makedirs(ckpt_dir)
"""------------------训练模型---------------------"""
train_batch_size = 128 # 训练集的mini_batch_size=128
test_batch_size = 256 # 测试集中调用的batch_size=256
epoches = 5 # 迭代周期
with tf.Session() as sess:
"""-------训练模型--------"""
# 初始化所有变量
tf.global_variables_initializer().run()
# 训练操作
for i in range(epoches):
train_batch = zip(range(0, len(trX), train_batch_size),range(train_batch_size, len(trX) + 1, train_batch_size))
for start, end in train_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})
# 每个周期用测试集中随机抽出test_batch_size个图片进行测试
test_indices = np.arange(len(teX)) # 返回一个array[0,1...len(teX)]
np.random.shuffle(test_indices) # 打乱这个array
test_indices = test_indices[0:test_batch_size]
# 获取测试集test_batch_size章图片的的预测结果
predict_result = sess.run(predict_op, feed_dict={X: teX[test_indices],p_keep_conv: 1.0,p_keep_hidden: 1.0})
# 获取真实的标签值
true_labels = np.argmax(teY[test_indices], axis=1)
# 计算准确率
accuracy = np.mean(true_labels == predict_result)
print("epoch", i, ":", accuracy)
# 保存模型
saver.save(sess,ckpt_dir+"/model.ckpt",global_step=i)
8.训练卷积神经网络+恢复模型
# 用Saver保存的模型的恢复
import tensorflow as tf
import numpy as np
import os
import cv2
# 屏蔽waring信息
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
"""------------------构建模型---------------------"""
# 定义输入输出的数据容器
X = tf.placeholder("float", [None, 28, 28, 1])
Y = tf.placeholder("float", [None, 10])
# 定义和初始化权重、dropout参数
def init_weights(shape):
return tf.Variable(tf.random_normal(shape, stddev=0.01))
w1 = init_weights([3, 3, 1, 32]) # 3X3的卷积核,获得32个特征
w2 = init_weights([3, 3, 32, 64]) # 3X3的卷积核,获得64个特征
w3 = init_weights([3, 3, 64, 128]) # 3X3的卷积核,获得128个特征
w4 = init_weights([128 * 4 * 4, 625]) # 从卷积层到全连层
w_o = init_weights([625, 10]) # 从全连层到输出层
p_keep_conv = tf.placeholder("float")
p_keep_hidden = tf.placeholder("float")
# 定义模型
def create_model(X, w1, w2, w3, w4, w_o, p_keep_conv, p_keep_hidden):
# 第一组卷积层和pooling层
conv1 = tf.nn.conv2d(X, w1, strides=[1, 1, 1, 1], padding='SAME')
conv1_out = tf.nn.relu(conv1)
pool1 = tf.nn.max_pool(conv1_out, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
pool1_out = tf.nn.dropout(pool1, p_keep_conv)
# 第二组卷积层和pooling层
conv2 = tf.nn.conv2d(pool1_out, w2, strides=[1, 1, 1, 1], padding='SAME')
conv2_out = tf.nn.relu(conv2)
pool2 = tf.nn.max_pool(conv2_out, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
pool2_out = tf.nn.dropout(pool2, p_keep_conv)
# 第三组卷积层和pooling层
conv3 = tf.nn.conv2d(pool2_out, w3, strides=[1, 1, 1, 1], padding='SAME')
conv3_out = tf.nn.relu(conv3)
pool3 = tf.nn.max_pool(conv3_out, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
pool3 = tf.reshape(pool3, [-1, w4.get_shape().as_list()[0]]) # 转化成一维的向量
pool3_out = tf.nn.dropout(pool3, p_keep_conv)
# 全连层
fully_layer = tf.matmul(pool3_out, w4)
fully_layer_out = tf.nn.relu(fully_layer)
fully_layer_out = tf.nn.dropout(fully_layer_out, p_keep_hidden)
# 输出层
out = tf.matmul(fully_layer_out, w_o)
return out
model = create_model(X, w1, w2, w3, w4, w_o, p_keep_conv, p_keep_hidden)
# 定义代价函数、训练方法、预测操作
cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=model, labels=Y))
train_op = tf.train.RMSPropOptimizer(0.001, 0.9).minimize(cost)
predict_op = tf.argmax(model, 1,name="predict")
# 定义一个saver
saver=tf.train.Saver()
# 定义存储路径
ckpt_dir="./ckpt_dir2"
with tf.Session() as sess:
"""-----加载模型,用导入的图片进行测试--------"""
# 载入图片
src = cv2.imread('img/5.png')
cv2.imshow("Pending pictures", src) # 待测图片
# 将图片转化为28*28的灰度图
src = cv2.cvtColor(src, cv2.COLOR_BGR2GRAY)
dst = cv2.resize(src, (28, 28), interpolation=cv2.INTER_CUBIC)
# 将灰度图转化为1*784的能够输入的网络的数组
picture = np.zeros((28, 28))
for i in range(0, 28):
for j in range(0, 28):
picture[i][j] = (255 - dst[i][j])
picture = picture.reshape(1, 28, 28, 1)
# 载入模型
saver.restore(sess,ckpt_dir+"/model.ckpt-1")
# 进行预测
predict_result = sess.run(predict_op, feed_dict={X: picture,p_keep_conv: 1.0,p_keep_hidden: 1.0})
print("你导入的图片是:",predict_result[0])
cv2.waitKey(0)
# 关闭会话
sess.close()