pytorch线性模型的建立

一、pytorch线性模型的建立

pytorch构建模型一般分为四块：1、数据集准备；2、设计模型；3、构建loss函数和优化器；4、训练

1.准备数据

在后面讲

2.构造模型 (源自于nn.Module)

class LinearModel(torch.nn.Module):
    # 在构建模型时必须有__init__()和forward()
    def __init__(self):
        # 调用父类的构造，必须要有
        super(LinearModel,self).__init__()
        # torch里面的一个类，Linear(in_features,out_features,bias),自动完成线性运算
        self.linear = torch.nn.Linear(1, 1)
        
    def forward(self, x):
        y_pred = self.linear(x)
        return y_pred
    # model是可调用的函数
 model = LinearModel()

3.构造loss函数和优化器(using PyTorch APT)

# MSELoss是继承自nn.module,MSELoss(size_average=true,reduce=true)
criterion = torch.nn.MSELoss(size_average=False)  # criterion需要y和y_hat
# BCEloss  cross-entropy  交叉熵
criterion = torch.nn.BCEloss(size_average=False) 
# 优化器
optimizer = torch.optim.SGD(model.parameters(), lr=0.01)

4.训练过程 (forward, backward, update)

# forawrd
for epoch in range(100):
    y_pred = model(x_data)   # 算出y_hat
    loss = criterion(y_pred, y_data)
    print(epoch, loss)
    
    optimizer.zero_grad()  #梯度归零
    loss.backward()  # 反向传播
    optimizer.step()  # 更新

5.最后打印出结果

#output weight and bias
# 加item
print('w=', model.linear.weight.item())
print('b = ', model.liear.bais.item())

6.测试阶段

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

7.添加sigmoid函数后的模型

import torch.nn.function as F
# F中有很多激活函数
class LogisticRegressionModel(torch.nn.Module):
    def __init__(self):
        super(LogisticRegressionModel,self).__init__()
        self.Linear = torch.nn.Linear(1, 1)
        
    def forward(self, x):
        # 在计算线性变换模型前添加一个sigmoid激活函数
        y_pred = F.sigmoid(self.linear(x))
        return(y_pred)

二、处理多维特征的输入

import torch
class Model(torch.nn.Module):
    def __init__(self):
        super(Model, self).___init__()
        self.linear1 = torch.nn.Linear(8, 6)
        self.linear2 = torch.nn.Linear(6, 4)
        self.linear3 = torch.nn.Linear(4, 1)
        self.sigmoid = torch.nn.Sigmoid()
        
        def forward(self, x):
            x = self.sigmoid(self.linear1(x))
            x = self.sigmoid(self.linear2(x))
            x = self.sigmoid(self.linear3(x))
            return x
        model = Model()
 
#构造loss函数和优化器
 criterion = torch.nn.BCEloss(size_average=False) 
 optimizer = torch.optim.SGD(model.parameters(), lr=0.01)
    
    # forawrd
for epoch in range(100):
    y_pred = model(x_data)   # 算出y_hat
    loss = criterion(y_pred, y_data)
    print(epoch, loss.item())
    # 反向传播
    optimizer.zero_grad()  #梯度归零
    loss.backward()  
    # 更新
    optimizer.step()  
    
    
    
    
    
    
    采用relu激活函数
    import torch
class Model(torch.nn.Module):
    def __init__(self):
        super(Model, self).___init__()
        self.linear1 = torch.nn.Linear(8, 6)
        self.linear2 = torch.nn.Linear(6, 4)
        self.linear3 = torch.nn.Linear(4, 1)
        self.activate = torch.nn.ReLU()
        self.sigmoid = torch.nn.Sigmoid()
        
        
        def forward(self, x):
            x = self.activate(self.linear1(x))
            x = self.activate(self.linear2(x))
            x = self.sigmoid(self.linear3(x))   # 最后输出激活函数改成sigmoid， 输出的曲线会比较光滑
            return x
        model = Model()