Saving and Loading Models



In [1]:

    
import numpy as np
import tensorflow as tf
import matplotlib.pyplot as plt
%matplotlib inline



In [2]:

    
np.random.seed(101)
tf.set_random_seed(101)

Full Network Example

Let's work on a regression example, we are trying to solve a very simple equation:

y = mx + b

y will be the y_labels and x is the x_data. We are trying to figure out the slope and the intercept for the line that best fits our data!

Artifical Data (Some Made Up Regression Data)



In [3]:

    
x_data = np.linspace(0, 10, 10) + np.random.uniform(-1.5, 1.5, 10)



In [4]:

    
x_data









    Out[4]:





array([0.04919588, 1.32311387, 0.8076449 , 2.3478983 , 5.00027539,
       6.55724614, 6.08756533, 8.95861702, 9.55352047, 9.06981686])



In [5]:

    
y_label = np.linspace(0, 10, 10) + np.random.uniform(-1.5, 1.5, 10)



In [6]:

    
plt.plot(x_data,
         y_label,
         '*')









    Out[6]:





[<matplotlib.lines.Line2D at 0x159c5f9fba8>]

Variables



In [7]:

    
np.random.rand(2)









    Out[7]:





array([0.68530633, 0.51786747])



In [8]:

    
m = tf.Variable(0.39)
b = tf.Variable(0.2)

Cost Function



In [9]:

    
error = tf.reduce_mean(y_label - (m * x_data + b))

Optimizer



In [10]:

    
optimizer = tf.train.GradientDescentOptimizer(learning_rate = 0.0005)
train = optimizer.minimize(error)

Initialize Variables



In [11]:

    
init = tf.global_variables_initializer()

Saving The Model



In [12]:

    
saver = tf.train.Saver()

Create Session and Run!



In [13]:

    
with tf.Session() as sess:
    
    sess.run(init)
    
    epochs = 300
    
    for i in range(epochs):
        
        sess.run(train)

    # Fetch Back Results
    final_slope , final_intercept = sess.run([m,b])
    
    # Save the checkpoint
    saver.save(sess,'new_models/my_second_model.ckpt')

Evaluate Results



In [14]:

    
x_test = np.linspace(-1, 11, 10)
y_pred_plot = final_slope * x_test + final_intercept

plt.plot(x_test, y_pred_plot,'r')

plt.plot(x_data, y_label,'*')









    Out[14]:





[<matplotlib.lines.Line2D at 0x159c60ae160>]

Loading a Model



In [15]:

    
with tf.Session() as sess:
    
    # Restore the model
    saver.restore(sess,'new_models/my_second_model.ckpt')
    

    # Fetch Back Results
    restored_slope, restored_intercept = sess.run([m,b])









    



INFO:tensorflow:Restoring parameters from new_models/my_second_model.ckpt



In [16]:

    
print(restored_slope)



In [17]:

    
print(restored_intercept)



In [18]:

    
x_test = np.linspace(-1, 11, 10)
y_pred_plot = restored_slope * x_test + restored_intercept

plt.plot(x_test,y_pred_plot,'r')

plt.plot(x_data,y_label,'*')









    Out[18]:





[<matplotlib.lines.Line2D at 0x15a716af550>]