Gradient Descent Algorithm 구현

import tensorflow as tf
 
x_data = [1,2,3]
y_data = [1,2,3]
 
W = tf.Variable(tf.random_normal([1]), name = 'weight')
 
X = tf.placeholder(tf.float32)
Y = tf.placeholder(tf.float32)
 
hypothesis = W * X
 
cost = tf.reduce_sum(tf.square(hypothesis -Y))
 
learning_rate = 0.1
gradient = tf.reduce_mean((W * X - Y) * X)
descent = W - learning_rate * gradient
update = W.assign(descent) 
 
sess = tf.Session()
 
sess.run(tf.global_variables_initializer())
 
for step in range(21):
    sess.run(update, feed_dict={X: x_data, Y: y_data})
    print(step, sess.run(cost, feed_dict{X: x_data, Y: y_data}), sess.run(W))
 
 

위 코드는 경사하강법 알고리즘을 구현한 예제이다.

optimizer = tf.train.GradientDesecdentOptimizer(learning rate = 0.1)
train = optimizer.minimize(cost)

물론 이렇게 쉽게 

learning_rate = 0.1
gradient = tf.reduce_mean((W * X - Y) * X)
descent = W - learning_rate * gradient
update = W.assign(descent)

이거 대신 사용 할 수 있다