《PyTorch深度學習實踐》完結合集_嗶哩嗶哩_bilibili

P5--用Pytorch實現線性迴歸

建立模型四大步驟

一、Prepare dataset

mini-batch：x、y必須是矩陣

## Prepare Dataset:mini-batch, X、Y是3X1的Tensor
x_data = torch.Tensor([[1.0], [2.0], [3.0]])
y_data = torch.Tensor([[2.0], [4.0], [6.0]])

二、Design model

1、重點是構造計算圖

##Design Model

##構造類，繼承torch.nn.Module類
 

class LinearModel(torch.nn.Module):
    ## 建構函式，初始化物件
    def __init__(self):
        ##super呼叫父類
        super(LinearModel, self).__init__()
        ##構造物件，Linear Unite，包含兩個Tensor：weight和bias，引數(1, 1)是w的維度
        self.linear = torch.nn.Linear(1, 1)
        
    ## 建構函式，前饋運算
    def forward(self, x):
        
 
## w*x+b
        y_pred = self.linear(x)
        return y_pred
    
model = LinearModel()

2、設定w的維度，後一層的神經元數量 X 前一層神經元數量

三、Construct Loss and Optimizer

##Construct Loss and Optimizer

##損失函式，傳入y和y_presd
criterion = torch.nn.MSELoss(size_average = False)

##優化器，model.parameters()找出模型所有的引數，Lr--學習率
 

optimizer = torch.optim.SGD(model.parameters(), lr=0.01)

1、損失函式

2、優化器

可用不同的優化器進行測試對比

四、Training cycle

## Training cycle

for epoch in range(100):
    y_pred = model(x_data)
    loss = criterion(y_pred, y_data)
    print(epoch, loss)
    
    ##梯度歸零
    optimizer.zero_grad()
    ##反向傳播
    loss.backward()
    ##更新
    optimizer.step()

完整程式碼

import torch

## Prepare Dataset:mini-batch, X、Y是3X1的Tensor
x_data = torch.Tensor([[1.0], [2.0], [3.0]])
y_data = torch.Tensor([[2.0], [4.0], [6.0]])


##Design Model

##構造類，繼承torch.nn.Module類
class LinearModel(torch.nn.Module):
    ## 建構函式，初始化物件
    def __init__(self):
        ##super呼叫父類
        super(LinearModel, self).__init__()
        ##構造物件，Linear Unite，包含兩個Tensor：weight和bias，引數(1, 1)是w的維度
        self.linear = torch.nn.Linear(1, 1)
        
    ## 建構函式，前饋運算
    def forward(self, x):
        ## w*x+b
        y_pred = self.linear(x)
        return y_pred
    
model = LinearModel()

##Construct Loss and Optimizer

##損失函式，傳入y和y_presd
criterion = torch.nn.MSELoss(size_average = False)

##優化器，model.parameters()找出模型所有的引數，Lr--學習率
optimizer = torch.optim.SGD(model.parameters(), lr=0.01)


## Training cycle

for epoch in range(100):
    y_pred = model(x_data)
    loss = criterion(y_pred, y_data)
    print(epoch, loss)
    
    ##梯度歸零
    optimizer.zero_grad()
    ##反向傳播
    loss.backward()
    ##更新
    optimizer.step()
    
## Outpue weigh and bias
print('w = ', model.linear.weight.item())
print('b = ', model.linear.bias.item())

## Test Model
x_test = torch.Tensor([[4.0]])
y_test = model(x_test)
print('y_pred = ', y_test.data)

執行結果

訓練100次後，得到的 weight and bias，還有預測的y