參考——莫煩python
包括：1.變數定義
2.session控制
3.佔位符
4.新增層
5.搭建神經網路
#!/usr/bin/env python
# -*- coding: utf-8 -*-
# @time : 18/2/27 下午5:11
# @author : cicada@hole
# @file : tf.py
# @desc : tensorflow測試檔案
# @link :
import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
def func1():
# dx_data = np.random.rand(100).astype(np.float32)
y_data = x_data * 0.1 + 0.3
# mweights = tf.variable(tf.random_uniform([1], -1.0, 1.0))
biases = tf.variable(tf.zeros([1]))
y = weights * x_data + biases
loss = tf.reduce_mean(tf.square(y - y_data))
optimizer = tf.train.gradientdescentoptimizer(0.5)
train = optimizer.minimize(loss)
init = tf.global_variables_initializer()
session = tf.session()
session.run(init)
for step in range(201):
session.run(train)
if step % 20 == 0:
print(step, session.run(weights), session.run(biases))
def sessioncontrol():
# 建立兩個矩陣
matrix1 = tf.constant([[3,3],[2,3]])
matrix2 = tf.constant([[2],[2]])
product = tf.matmul(matrix1, matrix2)
#用session來啟用product並得到計算結果
sess = tf.session()
# method1
result = sess.run(product)
print(result)
sess.close()
# method 2
with tf.session() as sess:
result2 = sess.run(product)
print(result2)
def tfvardefine():
# 定義變數 語法： state = tf.variable()
state = tf.variable(0, name='counter')
# 定義常量 one
one = tf.constant(1)
# 定義加法步驟（此步並沒有直接計算）
new_value = tf.add(state, one)
# 將state 更新為 new_value
update = tf.assign(state, new_value)
# 定義的變數需要啟用
init = tf.global_variables_initializer()
# 使用session
with tf.session() as sess:
sess.run(init)
for _ in range(3):
sess.run(update)
print('---------state ',sess.run(state)) # 需要把sess的指標指向state 再print
#-------佔位符placeholder----
def placeholdertest():
input1 = tf.placeholder(tf.float32) #大部分形式f32
input2 = tf.placeholder(tf.float32)
output = tf.multiply(input1, input2) #乘法運算
with tf.session() as sess:
print(sess.run(output, feed_dict=))
# feed_dict和placeholder進行繫結，能傳值
#----激勵函式-----如sigmod等
def activationfunctiontest():
pass
#----新增層 ----add_layer()----
def add_layer(inputs, in_size, out_size, activation_function=none):
weights = tf.variable(tf.random_normal([in_size, out_size])) #in_size行out_size列的隨機矩陣
biases = tf.variable(tf.zeros([1, out_size]) + 0.1) #biases推薦值不為零
wx_plus_b = tf.matmul(inputs, weights) + biases #神經網路未啟用的值
#activation_function = none時，輸出wx_plus_b,不為none時，輸出啟用後的wx_plus_b
if activation_function is none:
outputs = wx_plus_b
else:
outputs = activation_function(wx_plus_b)
return outputs
#----搭建網路-----
def buildnn():
#匯入資料
x_data = np.linspace(-1, 1, 300, dtype=np.float32)[:, np.newaxis]
noise = np.random.normal(0, 0.05, x_data.shape).astype(np.float32)
y_data = np.square(x_data) - 0.5 + noise
xs = tf.placeholder(tf.float32, [none, 1])
ys = tf.placeholder(tf.float32, [none, 1])
#輸入層1個、隱藏層10個、輸出層1個的神經網路
l1 = add_layer(xs, 1, 10, activation_function=tf.nn.relu)#輸入層只有乙個屬性
#輸入就是隱藏層的輸出——l1，輸入有10層（隱藏層的輸出層），輸出有1層
prediction = add_layer(l1, 10, 1, activation_function=none)
loss = tf.reduce_mean(tf.reduce_sum(tf.square(ys - prediction), reduction_indices=[1]))#平均誤差
# reduction_indices表示函式的處理維度
# 讓機器學習提公升準確率 以0.1的效率來最小化誤差loss
train_step = tf.train.gradientdescentoptimizer(0.1).minimize(loss)
init = tf.global_variables_initializer()
sess = tf.session()
sess.run(init)
fig = plt.figure()
ax = fig.add_subplot(1, 1, 1)
ax.scatter(x_data, y_data)
plt.ion() # 連續顯示
plt.show()
for i in range(1000):
sess.run(train_step, feed_dict=)
if i % 50 == 0:
# to visualize the result and improvement
try:
ax.lines.remove(lines[0])
except exception:
pass
prediction_value = sess.run(prediction, feed_dict=)
# plot the prediction
lines = ax.plot(x_data, prediction_value, 'r-', lw=5) #顯示**的資料
tensorflow 1 共享變數
共享變數 reuse variables example1 with tf.variable scope try 先建立兩個變數w1，w2 w2 tf.get variable w1 shape 2,3,4 dtype tf.float32 w3 tf.get variable w2 shape 2...
tensorflow1 構建線性模型
x是給定的輸入資料 使用tensorflow構建乙個模型，開始的時候，w和b全部給成0，讓其訓練，使其接近預設的模型。即讓w接近0.1，b接近0.2 import tensorflow as tf import numpy as np x data np.random.rand 100 y data...
神經網路 tensorflow 1
import tensorflow as tf import numpy as np create data x data np.random.rand 100 astype np.float32 在tensorflow中大部分的資料的資料型別都是float32 y data x data 0.1 ...