In [1]:

    

import tensorflow as tf


    

In [3]:

    

import numpy as np


    

In [5]:

    

x_data = np.random.rand(100).astype(np.float32)
y_data = x_data * 0.1 + 00.3


    

In [6]:

    

weight = tf.Variable(tf.random_uniform([1], -1.0, 1.0))


    

In [9]:

    

biases = tf.Variable(tf.zeros([1]))


    

In [10]:

    

y = weight * x_data + biases


    

In [11]:

    

loss = tf.reduce_mean(tf.square(y-y_data))


    

In [12]:

    

optimizer = tf.train.GradientDescentOptimizer(0.5)


    

In [13]:

    

train = optimizer.minimize(loss)


    

In [16]:

    

init = tf.global_variables_initializer()


    

In [17]:

    

sess = tf.Session()


    

In [18]:

    

sess.run(init)


    

In [21]:

    

for step in range(201):
    sess.run(train)
    if step  % 20 == 0:
        print(step, sess.run(weight), sess.run(biases))


    

    




0 [ 0.69158053] [-0.0334721]
20 [ 0.27189481] [ 0.20404713]
40 [ 0.14983082] [ 0.2721841]
60 [ 0.11444556] [ 0.2919364]
80 [ 0.10418765] [ 0.29766244]
100 [ 0.10121397] [ 0.29932237]
120 [ 0.10035191] [ 0.29980358]
140 [ 0.10010203] [ 0.29994306]
160 [ 0.10002957] [ 0.2999835]
180 [ 0.10000858] [ 0.29999521]
200 [ 0.10000247] [ 0.29999864]

In [ ]:

    




    

In [ ]:

    

def g(n):


    

In [ ]:

    




    

In [ ]:

    




    

In [ ]:

    




    

In [ ]:

    




    

In [ ]:

    




    

In [ ]:

    




    

In [ ]:

    




    

In [ ]:

    




    

In [ ]:

    




    

In [ ]:

    




    

In [ ]: