Lec-03 Liner Regression and How to minimize cost LAB

Cost function in pure Python

import numpy as np

X = np.array([1, 2, 3])
Y = np.array([1, 2, 3])    

def cost_func(W, X, Y):
    c = 0
    for i in range(len(X)):
        c += (W * X[i] - Y[i]) ** 2
    return c / len(x)	# 편차 제곱의 평균
for feed_W in np.linspace(-3, 5, num=15):	# -3~5까지 15개의 구간
    curr_cost = cost_func(feed_W, X, Y)
    print("{:6.3f} | {:10.5f}".format(feed_W, curr_cost))

Cost function in TensorFlow

import numpy as np
    
X = np.array([1, 2, 3])
Y = np.array([1, 2, 3])    
    
def cost_func(W, X, Y):
    hypothesis = X * W
    return tf.reduce_mean(tf.square(hypothesis - Y))
            
W_values = np.linspace(-3, 5, num=15)
cost_values = []
    
for feed_W in W_values:
    curr_cost = cost_func(feed_W, X, Y)
    cost_values.append(curr_cost)
    print("{:6.3f} | {:10.5f}".format(feed_W, curr_cost))

Gradient descent

tf.random.set_seed(0)	# for reproducibility
    
X = [1., 2., 3., 4.]
Y = [1., 2., 3., 4.]
    
W = tf.Variable(tf.random_normal([1], -100. 100.))	# random
# W = tf.Variable([5.0]) 
# W에 어떤 값을 주어도 cost는 0으로 W는 특정한 값으로 수렴.
    
for step in range(300):
    hypothesis = W * X
    cost = tf.reduce_mean(tf.square(hypothesis - Y))
            
    alpha = 0.01
    gradient = tf.reduce_mean(tf.multiply(tf.multiply(W, X) - Y, X))
    descent = W - tf.multiply(alpha, gradient)	# 새로운 W값
    W.assign(descent)
            
    if step % 10 == 0:
        print('{:5} | {:10.4f} | {:10.6f}'.format(step, cost.numpy(), W.numpy()[0]))

모두를 위한 딥러닝 시즌2 - Tensorflow