[转载] TensorFlow2.0 学习 线性回归

参考链接： 使用Tensorflow进行线性回归

import numpy as np

import tensorflow as tf

 

X = np.array([1, 2, 3, 4, 5], dtype=np.float32)

y = np.array([10, 19, 31, 42, 53], dtype=np.float32)

 

# X = (X_raw - X_raw.min()) / (X_raw.max() - X_raw.min())

# y = (y_raw - y_raw.min()) / (y_raw.max() - y_raw.min())

 

 

# X = tf.constant(X)

# y = tf.constant(y)

 

w = tf.Variable(initial_value=0.)

b = tf.Variable(initial_value=0.)

variables = [w, b]

 

num_epoch = 10000

optimizer = tf.optimizers.SGD(learning_rate=1e-3) # 声明了一个梯度下降优化器optimizer,学习率为1e-3

for e in range(num_epoch):

    # 使用tf.GradientTape()记录损失函数的梯度信息

    with tf.GradientTape() as tape:

        y_pred = w * X + b

        loss = tf.reduce_mean(tf.square(y_pred - y))

    # TensorFlow自动计算损失函数关于自变量（模型参数）的梯度

    grads = tape.gradient(loss, variables)

 

    # TensorFlow自动根据梯度更新参数

    optimizer.apply_gradients(grads_and_vars=zip(grads, variables))

 

 

 

 

print(w, b)

 

print(float(w))

10.8587007522583

print(b.numpy())

-1.5509015

print(w.numpy())

10.858701