48、tensorflow入门二,线性模型的拟合
2017-10-13 11:11
405 查看
import tensorflow as tf import numpy as np #生成2维的100个0-1的随机数 x_data = np.float32(np.random.rand(2,100)) 计算内积,x_data一共100个值,每个值都是2维的向量,用[0.1,0.2]和每一个向量计算数量积,然后加起来
x_data=
[[ 0.27559635 0.35930911 0.77030689 0.71818703 0.49122271 0.43190494
0.644674 0.32078174 0.64300877 0.81156862 0.30409896 0.56623858
0.97026539 0.60468578 0.34658566 0.86189109 0.5117926 0.46588144
0.27591956 0.49946061 0.47744861 0.80954593 0.10624354 0.80820572
0.24388497 0.75275064 0.11153043 0.01847375 0.70894343 0.86511648
0.05999189 0.94242656 0.35399687 0.53131646 0.80690706 0.28856653
0.2685678 0.86655128 0.49340782 0.84330899 0.26634833 0.94808429
0.32813659 0.60548925 0.37914801 0.93819922 0.16300483 0.28346273
0.25481561 0.59326059 0.64435166 0.71002674 0.47835174 0.16478723
0.83618289 0.89197201 0.77212745 0.83254766 0.1176443 0.45999372
0.17508474 0.99125117 0.19204263 0.88548642 0.16025347 0.58622926
0.14167576 0.6784007 0.77483946 0.90998834 0.79064935 0.76124579
0.11255023 0.63665706 0.84133714 0.01388079 0.7478959 0.34365693
0.66228282 0.56429321 0.97419363 0.46427366 0.71639329 0.67420006
0.72603422 0.35240087 0.43481046 0.04478104 0.83910578 0.03852031
0.63502115 0.54327303 0.05116724 0.75744271 0.23107423 0.25379685
0.1543453 0.65174055 0.60880935 0.41448417]
[ 0.53852242 0.24170624 0.51843584 0.41295227 0.22256 0.05581184
0.42046902 0.49984431 0.83284378 0.44403863 0.43289256 0.04277489
0.97858369 0.32724616 0.69393569 0.80431139 0.20591183 0.10109164
0.07850602 0.61202133 0.04476574 0.99151891 0.1714984 0.83303201
0.1769124 0.97038633 0.71856993 0.97560126 0.80447757 0.48544171
0.95302963 0.21392477 0.72407377 0.32749009 0.87037027 0.23632777
0.09344739 0.49172315 0.27751547 0.3205907 0.42732051 0.0938397
0.65851027 0.75118226 0.30484736 0.69336241 0.96847987 0.08743072
0.5060789 0.128803 0.12509818 0.77400607 0.99729323 0.25656971
0.28877217 0.26310787 0.22661451 0.38361222 0.64689898 0.26246113
0.41836309 0.96913052 0.34863174 0.26865834 0.96321774 0.02932074
0.51096094 0.93037766 0.3862699 0.77660888 0.50103205 0.35242727
0.96469277 0.71796703 0.90261179 0.9502635 0.2554118 0.41087386
0.13807607 0.10848427 0.27238116 0.81126028 0.35296583 0.5478636
0.35726911 0.54948765 0.3683508 0.57419771 0.28177765 0.80673724
0.14792147 0.5277251 0.17725706 0.5937981 0.86571193 0.09969555
0.92383957 0.95939624 0.76108253 0.30095646]]
y_data =
[ 0.43526412 0.38427216 0.48071786 0.45440916 0.39363427 0.35435286
0.4485612 0.43204704 0.53086963 0.46996459 0.41698841 0.36517884
0.59274328 0.42591781 0.4734457 0.54705139 0.39236163 0.36680647
0.34329316 0.47235033 0.35669801 0.57925838 0.34492403 0.54742697
0.35977098 0.56935233 0.45486703 0.49696763 0.53178986 0.48359999
0.49660512 0.43702761 0.48021444 0.41862966 0.55476476 0.37612221
0.34554626 0.48499976 0.40484388 0.44844904 0.41209894 0.41357637
0.46451571 0.51078538 0.39888427 0.53249241 0.50999646 0.34583242
0.42669734 0.38508666 0.3894548 0.52580389 0.54729382 0.36779267
0.44137272 0.44181877 0.42253565 0.45997721 0.44114423 0.3984916
0.40118109 0.59295122 0.38893061 0.44228031 0.50866889 0.36448708
0.41635976 0.5539156 0.45473793 0.54632061 0.47927135 0.44661003
0.50419358 0.50725911 0.56465607 0.49144078 0.42587195 0.41654046
0.3938435 0.37812617 0.45189559 0.50867942 0.4422325 0.47699273
0.44405724 0.44513762 0.41715121 0.41931765 0.44026611 0.46519948
0.39308641 0.45987232 0.34056814 0.49450389 0.49624981 0.34531879
0.50020244 0.5570533 0.51309744 0.40163971]
y_data = np.dot([0.100,0.200],x_data) + 0.300 b = tf.Variable(tf.zeros([1])) W = tf.Variable(tf.random_uniform([1,2],-1.0,1.0)) y = tf.matmul(W,x_data) + b loss = tf.reduce_mean(tf.square(y - y_data)) optimizer = tf.train.GradientDescentOptimizer(0.5) train = optimizer.minimize(loss) init = tf.global_variables_initializer() sess = tf.Session() sess.run(init) for step in np.arange(0,201): sess.run(train) if step % 20 == 0: print(step,sess.run(W),sess.run(b)) 结果如下所示
>>> import testTensorflow
0 [[ 0.61865866 0.54848659]] [-0.35512698]
20 [[ 0.27226886 0.31722128]] [ 0.14856057]
40 [[ 0.15020847 0.23340638]] [ 0.25623968]
60 [[ 0.11460328 0.20955895]] [ 0.28735051]
80 [[ 0.10424114 0.20274341]] [ 0.29634258]
100 [[ 0.10123044 0.20078911]] [ 0.2989423]
120 [[ 0.10035671 0.20022736]] [ 0.29969406]
140 [[ 0.10010336 0.20006558]] [ 0.2999115]
160 [[ 0.10002995 0.20001893]] [ 0.29997438]
180 [[ 0.10000868 0.20000547]] [ 0.29999259]
200 [[ 0.10000249 0.20000155]] [ 0.29999787]
>>>
相关文章推荐
- 48、tensorflow入门二,线性模型的拟合
- tensorflow入门(2)线性模型
- 机器学习排序LTR入门——线性模型
- tensorflow线性模型以及Wide deep learning
- 【机器学习】Tensorflow概率编程:线性混合模型
- 79、tensorflow计算一个五层神经网络的正则化损失系数、防止网络过拟合、正则化的思想就是在损失函数中加入刻画模型复杂程度的指标
- 二、tensorflow线性拟合实例。
- 79、tensorflow计算一个五层神经网络的正则化损失系数、防止网络过拟合、正则化的思想就是在损失函数中加入刻画模型复杂程度的指标
- TensorFlow 入门 3 ——变量管理和模型持久化
- 使用TensorFlow实现RNN模型入门篇1
- 简单Tensorflow线性拟合类及tf.get_variable()使用示例
- 深度学习笔记——深度学习框架TensorFlow(五)[TensorFlow大规模线性模型教程]
- Tensorflow学习教程------简单练一波,线性模型
- [机器学习入门] 李宏毅机器学习笔记-25(Sturctured Linear Model;结构化预测-线性模型)
- Tensorflow,线性模型,梯度下降
- tensorflow入门Day1-简单线性回归
- 使用TensorFlow实现RNN模型入门篇
- 使用TensorFlow实现RNN模型入门篇2--char-rnn语言建模模型
- sklearn 线性模型使用入门
- 深度学习入门实战(二)-用TensorFlow训练线性回归