【PyTorch深度学习实践 P2】y=wx+b

阿斯蒂芬幻觉

于 2024-09-02 09:30:34 发布

阅读量300

点赞数 4

分类专栏：刘二大人PyTorch 文章标签：深度学习 pytorch 人工智能

本文链接：https://blog.csdn.net/weixin_45665700/article/details/141808168

版权

刘二大人PyTorch 专栏收录该内容

4 篇文章 0 订阅

订阅专栏

模型y=wx+b的实现，及绘图

# FILE: 学习深度学习/Linear_Model
# USER: mcfly
# IDE: PyCharm
# CREATE TIME: 2024/9/2 8:29
# DESCRIPTION: 刘二大人 线性模型

import numpy as np
from matplotlib import pyplot as plt
from mpl_toolkits.mplot3d import Axes3D

x_train = [1.0, 2.0, 3.0]
y_train = [3.0, 5.0, 7.0]

def forward( w, b, x):
    return w * x + b

def loss( y, y_head=0.0):
    return (y-y_head) * (y-y_head)

y_heads = []
mse_dict = {}

for w in np.arange(0.0,4.1,0.1):
    y_heads_i = []
    for b in np.arange( 0.0, 2.1, 0.1):
        loss_sum = 0.0
        for xi, yi in zip( x_train, y_train ):
            y_head_i = forward(w, b, xi)
            y_heads_i.append( y_head_i )
            loss_sum += loss( yi, y_head_i )
        loss_sum /= len(x_train)
        mse_dict[(w,b)] = loss_sum
        print( "w:{0}\nb:{1}\ny:{2}\nloss:{3}".format(w, b, y_heads_i, loss_sum))
    y_heads.append(y_heads_i)

local_min = min(mse_dict, key=lambda x:mse_dict[x])
print( "{} is the w that minimize the loss.".format(local_min))



x = [key[0] for key in mse_dict.keys()]
y = [key[1] for key in mse_dict.keys()]
z = list(mse_dict.values())

# 三维散点图
fig = plt.figure(num="Loss") 
ax = fig.add_subplot(111, projection='3d')
ax.scatter(x, y, z, c='b', marker='o')
# 画出坐标轴
ax.set_xlabel('w') 
ax.set_ylabel('b')
ax.set_zlabel('loss')

plt.show()