PyTorch深度学习实践03

梯度下降

-- codeing = utf-8 --

@Time :2021/4/12 20:25

@Author:sueong

@File:03.py

@Software:PyCharm

import numpy as np
import matplotlib.pyplot as plt


x_data=[1.0,2.0,3.0]
y_data=[2.0,4.0,6.0]
w=1.0#初始值假设为1.0 w=w-a*cost对w的偏导

cost_list=[]
epoch_list=[]
def forword(x):
    return x*w



def cost(xs,ys):
    l_sum = 1
    for x,y in zip(x_data,y_data):
        y_pre = forword(x)
        l_sum+=(y_pre-y)**2
    return l_sum/len(xs)

def gradient(xs,ys):
    grad=0
    for x, y in zip(x_data, y_data):
        y_pre = forword(x)
        grad+=2*x*(y_pre-y)
    return grad/len(xs)

print('predict before traing',4,forword(4))
for epoch in range(100):
    epoch_list.append(epoch)
    cost_val=cost(x_data,y_data)
    cost_list.append(cost_val)
    grad_val=gradient(x_data,y_data)
    w-=0.01*grad_val

    print('Epoch=',epoch,'w=',w,'loss',cost_val)
print('predict(after training)',4,forword(4))



plt.plot(epoch_list,cost_list)
plt.ylabel('epoch')
plt.xlabel('cost')
plt.show()

随机梯度下降

从左边变成右边这样

# -*- codeing = utf-8 -*-
# @Time :2021/4/12 20:58
# @Author:sueong
# @File:Stochastic Gradient Descent.py
# @Software:PyCharm
# -*- codeing = utf-8 -*-
# @Time :2021/4/12 20:25
# @Author:sueong
# @File:03.py
# @Software:PyCharm
import numpy as np
import matplotlib.pyplot as plt


x_data=[1.0,2.0,3.0]
y_data=[2.0,4.0,6.0]
w=1.0#初始值假设为1.0 w=w-a*cost对w的偏导

cost_list=[]
epoch_list=[]
def forword(x):
    return x*w



def loss(x,y):
    y_pre = forword(x)
    return (y_pre-y)**2

def gradient(x,y):
    return 2*x*(x*w-y)

print('predict before traing',4,forword(4))
for epoch in range(100):
    epoch_list.append(epoch)
    for x,y in zip(x_data,y_data):
        grad=gradient(x,y)#对每个样本grad就更新w
        w=w-0.01*grad
        print('\tgrad：',x,y,grad)
        l=loss(x,y)

    print('Epoch=',epoch,'w=',w,'loss',l)
    cost_list.append(l)

print('predict(after training)',4,forword(4))



plt.plot(epoch_list,cost_list)
plt.ylabel('epoch')
plt.xlabel('cost')
plt.show()