Second part of code for batch training

Batch-training-2.py

loss = tf.reduce_mean(tf.square(Y_pred - Y_target))

# Declare the optimizer (G.D)
my_opt = tf.train.GradientDescentOptimizer(0.02)
train_step = my_opt.minimize(loss)

loss_batch = []

for i in range(100):
    #pick a random 20 data points
    rand_index = np.random.choice(100, size=batch_size)
    x_batch = np.transpose([x_vals[rand_index]])  # Transpose to the correct shape
    y_batch = np.transpose([y_vals[rand_index]])
    sess.run(train_step, feed_dict={X_data: x_batch, Y_target:y_batch})
    #Print the result after 5 intervals
    if(i+1) % 5 == 0:
        print('Step #', str(i+1), 'W = ', str(sess.run(W)))
        temp_loss = sess.run(loss, feed_dict={X_data: x_batch, Y_target:y_batch})
        loss_batch.append(temp_loss)
        print('Loss = ', temp_loss)