3. 梯度下降

B站 刘二大人老师的课程代码
PyTorch 深度学习实践

梯度下降

import torch
import numpy as np
import matplotlib.pyplot as plt

#样本数据
x_data=[1.,2.,3.] #输入样本
y_data=[4.,5.,6.] #输出样本

w=1
#定义模型
def forward(x):
    return x*w

#定义损失函数
def cost(xs,ys):
    cost=0
    for x,y in zip(xs,ys):
        y_pred=forward(x)
        cost+=(y_pred-y)**2
    return cost/len(xs)

#定义梯度

def grident(xs,ys):
    grad=0
    for x,y in zip(xs,ys):
        grad+=2*x*(x*w-y)
    return grad/len(xs)

print('Predict (before training)',4,forward(4))


epoch_list=[] #存储损失函数数据
cost_list=[]

for epoch in range(100):
    cost_val=cost(x_data,y_data)
    grad_val=grident(x_data,y_data)
    w-=0.01*grad_val
    print('epoch:',epoch,'w=',w,'loss=',cost_val)
    epoch_list.append([epoch])
    cost_list.append([cost_val]) #添加每次计算损失函数的值
      
print('Predict (after training)',4,forward(4))

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

随机梯度下降

更新每个样本的梯度

import torch
import numpy as np
import matplotlib.pyplot as plt

# -- coding: utf-8 --
# 样本数据
x_data = [1., 2., 3.]  # 输入样本
y_data = [4., 5., 6.]  # 输出样本

w = 1

# 定义模型
def forward(x):
    return x * w

# 定义损失函数
def loss(x, y):
    y_pred = forward(x)
    return (y_pred-y)**2


# 定义梯度

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


print('Predict (before training)', 4, forward(4))

epoch_list = []  # 存储损失函数数据
loss_list = []

for epoch in range(100):
    grad_val = gradient(x, y)
    w -= 0.01 * grad_val
    print("\tgrad:",x,y,grad)
    
    l = loss(x, y)
    print('progress:', epoch, 'w=', w, 'loss=', l)
    
    epoch_list.append([epoch])
    loss_list.append([l])  # 添加每次计算损失函数的值

print('Predict (after training)', 4, forward(4))

plt.plot(epoch_list, loss_list)
plt.xlabel('epoch')
plt.ylabel('loss')
plt.show()