end code for stochastic training on python using Tensorflow

stochastic_training-2.py

# Declare an optimizer: here i use gradient descent
my_opt = tf.train.GradientDescentOptimizer(learning_rate=0.02)

#Create the train step
train_step = my_opt.minimize(loss)

n_iterations = 100

loss_stochastic = []

for i in range(n_iterations):
    rand_index = np.random.choice(100)
    rand_x = [x_val[rand_index]]
    rand_y = [y_val[rand_index]]
    #Run the graph
    sess.run(train_step, feed_dict={x_data: rand_x, y_target: rand_y})
   #Print the result after 5 intervals
    if(i+1) % 5 == 0:
        print('Step #', str(i+1), 'W = ', str(sess.run(W_st)))
        temp_loss = sess.run(loss, feed_dict={x_data: rand_x, y_target: rand_y})
        loss_stochastic.append(temp_loss)
        print('Loss = ', temp_loss)