由於語法渲染問題而影響閱讀體驗， 請移步博客閱讀～
本文GitPage地址

Tensorflow

  1. ##!/usr/locol/bin/python3.6
  3. import tensorflow as tf
  4. import numpy as np
  6. ## creat data
  7. x_data = np.random.rand(100).astype(np.float32)
  8. y_data = x_data*0.1 + 0.3
  11. #### creat tnsorflow structure start
  12. Weights = tf.Variable(tf.random_uniform([1],-1.0,1.0))
  13. biases = tf.Variable(tf.zeros([1]))
  15. y = Weightess =tf.Session()
  16. *x_data + biases
  18. loss = tf.reduce_mean(tf.square(y - y_data))
  19. optimizer = tf.train.GradientDescentOptimizer(0.5)
  21. train = optimizer.minimize(loss)
  23. init = tf.initialize_all_variables()
  25. ###create tensorflow structure end ###
  26. sess =tf.Session()
  27. sess.run(init)
  29. for step in range(201):
  30.  sess.run(train)
  31.  if step % 20 == 0:
  32.   print(step, sess.run(Weights),sess.run(biases))
  38. #### add a laier###
  40. def add_layer(inputs, in_size, out_size, activation_function=None):
  41.   Weights = tf.Variable(tf.random_normal([in_size, out_size]))
  42.   biases = tf.Variable(tf.zeros([1, out_size]) + 0.1)
  43.   Wx_plus_b = tf.matmul(inputs, Weights) + biases
  44.   if activation_function is None:
  45.      outputs = Wx_plus_b
  46.   else:
  47.      outputs = activation_function(Wx_plus_b)
  48.   return outputs
  55. ##################
  56. ###From  https://tensorflow.google.cn/get_started/premade_estimators##
  57. #############

Enjoy~

本文由Python腳本GitHub/語雀自動更新

由於語法渲染問題而影響閱讀體驗， 請移步博客閱讀～
本文GitPage地址

GitHub: Karobben
Blog:Karobben
BiliBili:史上最不正經的生物狗