tensorflow--建立一个简单的小网络

In [19]:

 

 

 

 

 

import tensorflow as tf

import numpy as np

# #简单的数据形网络

# #定义输入参数

# X=tf.constant(value=[[0.7,0.5]])

# W1=tf.Variable(tf.truncated_normal(shape=[2,3],mean=0,stddev=2))

# W2=tf.Variable(tf.truncated_normal(shape=[3,1],mean=0,stddev=2))

​

# #定义前向传播过程

# a=tf.matmul(X,W1)

# y=tf.matmul(a,W2)

​

# #变量初始化

# init=tf.global_variables_initializer()

​

# with tf.Session() as sess:

#     sess.run(init)

#     print("Y",sess.run(y))

​

#multiply是对应元素相乘,matul是矩阵相乘

# 定义网络变量

#s输入常亮

x_input=tf.placeholder(dtype=tf.float32,shape=[1,2],name="x_input")

#权重变量

W1=tf.Variable(tf.random_normal(shape=[2,3],mean=0,stddev=2))

W2=tf.Variable(tf.random_normal(shape=[3,1],mean=0,stddev=2))

​

​

#定义网络运算图

a=tf.matmul(x_input,W1)

Y=tf.matmul(a,W2)

​

#初始化全局变量

init=tf.global_variables_initializer()

​

​

​

with tf.Session() as sess:

    sess.run(init)

    result=sess.run(Y,feed_dict={x_input:[[0.7,0.5]]})

    print("result", result)

 

 

 

result [[-4.5023837]]


#在pycharm运行程序

import tensorflow as tf
import numpy as np


BATCH_SIZE = 8  # 一次输入网络的数据，称为batch。一次不能喂太多数据
SEED = 23455    # 产生统一的随机数


rdm = np.random.RandomState(SEED)
X = rdm.rand(32, 2)

i=0
Y=np.zeros((32,), dtype=np.int)
Y_=np.transpose([Y])
for (x0, x1) in X:
    y=int(x0 + x1 < 1)
    Y_[[i]]=y
    i=i+1

print("X:\n", X)
print("Y_:\n",Y_)

 # 1定义神经网络的输入、参数和输出,定义前向传播过程。
x = tf.placeholder(tf.float32, shape=(None, 2))
y_ = tf.placeholder(tf.float32, shape=(None, 1))

w1 = tf.Variable(tf.random_normal([2, 3], stddev=1, seed=1))
w2 = tf.Variable(tf.random_normal([3, 1], stddev=1, seed=1))

a = tf.matmul(x, w1)
y = tf.matmul(a, w2)


# 2定义损失函数及反向传播方法。
loss = tf.reduce_mean(tf.square(y - y_))
train_step = tf.train.GradientDescentOptimizer(0.001).minimize(loss)  # 三种优化方法选择一个就可以


with tf.Session() as sess:
    init = tf.global_variables_initializer()
    sess.run(init)
    # 输出目前（未经训练）的参数取值。
    print("w1:\n", sess.run(w1))
    print("w2:\n", sess.run(w2))
    print("\n")
    STEPS = 3000
    for i in range(STEPS):               #0-2999
        start = (i * BATCH_SIZE) % 32
        end = start + BATCH_SIZE
        sess.run(train_step, feed_dict={x: X[start:end], y_: Y_[start:end]})
        if i % 500 == 0:
            total_loss = sess.run(loss, feed_dict={x: X, y_: Y_})
            print("After %d training step(s), loss on all data is %g"%(i,total_loss))
    print("\n")
    print("w1:\n", sess.run(w1))
    print("w2:\n", sess.run(w2))

#比较完整的一个网络

import tensorflow as tf
import numpy as np
batch_size=8
seed=23455

#制造一些假数据
rng=np.random.RandomState(seed)
X=rng.rand(32,2)
print(X)

Y=np.zeros(shape=(32,1),dtype=np.int)

Y=[[np.int(x0+x1<1)]for (x0,x1) in X]

print(Y)

#定义网络
x_input=tf.placeholder(shape=[None,2],dtype=np.float,name="input")
y_output=tf.placeholder(shape=[None,1],dtype=np.float,name="output")

#定义变量
W1=tf.Variable(tf.random_normal(shape=[2,3],stddev=1,seed=1))
W2=tf.Variable(tf.random_normal(shape=[3,1],stddev=1,seed=1))

#定义静态网络函数
a=tf.matmul(x_input,W1)
y=tf.matmul(a,W2)

#定义损失函数
loss=tf.reduce_mean(tf.square(y-y_output))
train_step=tf.train.GradientDescentOptimizer(0.001).minimize(loss)
#初始化变量
init=tf.global_variables_initializer()

with tf.Session() as sess:
sess.run(init)
print("W1:\n",W1)
#开始批量载入数据
for i in range(2000):
data_start=(i*batch_size)%32
data_end=data_start+batch_size
#开始训练数据
sess.run(train_step,feed_dict={x_input:X[data_start:data_end],y_output:Y[data_start:data_end]})
#每隔一段就打印出损失值
if ((i)%500==0):
Loss=sess.run(loss,feed_dict={x_input:X,y_output:Y})
print("loss",Loss)
print("w1",sess.run(W1))
print("W2",sess.run(W2))

转载于:https://www.cnblogs.com/shuimuqingyang/p/11019308.html