深度学习-tensorflow-梯度计算和mes激活函数

import tensorflow as tf
import numpy as np
from sklearn.datasets import fetch_california_housing
from sklearn.preprocessing import StandardScaler

epochs = 1000000
learn_rate = 0.03

house = fetch_california_housing(data_home="/home/botter/data_sets", download_if_missing=True)
m, n = house.data.shape
# m row data, n column data
print(m, n)
print("**************************house.data:")
print(house.data)
print("**************************house.target:")
print(house.target)
#  add one column of value is 1
house_data_plus_bias = np.c_[np.ones((m, 1)), house.data]

# standardScaler
scaler = StandardScaler().fit(house_data_plus_bias)
print("**************************scaler:")
print(scaler)
scaler_house_data_plus_bias = scaler.transform(house_data_plus_bias)
print("**************************scaler_house_data_plus_bias:")
print(scaler_house_data_plus_bias)

X = tf.constant(scaler_house_data_plus_bias, dtype=tf.float32, name="X")
y = tf.constant(house.target.reshape(-1, 1), dtype=tf.float32, name="y")
print("**************************X:")
print(X)
print("**************************y")
print(y)
theta = tf.Variable(tf.random_uniform([n + 1, 1], -1.0, 1.0), name="theta")
print("***************************theta:")
print(theta)
y_predict = tf.matmul(X, theta, name="predictions")
print("***************************predictions")
print(y_predict)
error = y_predict - y
print("***************************error")
print(error)
mse = tf.reduce_mean(tf.square(error), name="mse")
print("***************************mes")
print(mse)
# 梯度的公式：(y_pred - y) * xj
gradients = 2/m * tf.matmul(tf.transpose(X), error)
training_op = tf.assign(theta, theta - learn_rate * gradients)

init = tf.global_variables_initializer()

with tf.Session() as sess:
    sess.run(init)

    for epoch in range(epochs):
        if epoch % 100 == 0:
            print("Epoch", epoch, "MSE = ", mse.eval())
        sess.run(training_op)

    best_theta = theta.eval()
    print(best_theta)