使用Pytorch实现自定义的交叉熵损失函数，对手写数字数据集进行分类

使用torch.nn自带损失函数

import torch
import torch.nn as nn
import matplotlib.pyplot as plt
import torchvision
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

torchvision.datasets.MNIST(r"D:\365天深度学习100列\Pytorch实战  第P1周：实现mnist手写数字识别", train=True, transform=None, target_transform=None, download=True)
train_ds = torchvision.datasets.MNIST('data', 
                                      train=True, 
                                      transform=torchvision.transforms.ToTensor(), # 将数据类型转化为Tensor
                                      download=True)

test_ds  = torchvision.datasets.MNIST('data', 
                                      train=False, 
                                      transform=torchvision.transforms.ToTensor(), # 将数据类型转化为Tensor
                                      download=True)



batch_size = 32
train_dl = torch.utils.data.DataLoader(train_ds, 
                                       batch_size=batch_size, 
                                       shuffle=True)

test_dl  = torch.utils.data.DataLoader(test_ds, 
                                       batch_size=batch_size)


imgs, labels = next(iter(train_dl))
print(imgs.shape)


import numpy as np


import torch.nn.functional as F

num_classes = 10  # 图片的类别数

class Model(nn.Module):
     def __init__(self):
        super().__init__()
         # 特征提取网络
        self.conv1 = nn.Conv2d(1, 32, kernel_size=3)  # 第一层卷积,卷积核大小为3*3
        self.pool1 = nn.MaxPool2d(2)                  # 设置池化层，池化核大小为2*2
        self.conv2 = nn.Conv2d(32, 64, kernel_size=3) # 第二层卷积,卷积核大小为3*3   
        self.pool2 = nn.MaxPool2d(2) 
                                      
        # 分类网络
        self.fc1 = nn.Linear(1600, 64)          
        self.fc2 = nn.Linear(64, num_classes)
     # 前向传播
     def forward(self, x):
        x = self.pool1(F.relu(self.conv1(x)))     
        x = self.pool2(F.relu(self.conv2(x)))

        x = torch.flatten(x, start_dim=1)

        x = F.relu(self.fc1(x))
        x = self.fc2(x)
       
        return x


from torchinfo import summary
# 将模型转移到GPU中（我们模型运行均在GPU中进行）
model = Model().to(device)

summary(model)


# 三、 训练模型
# 1. 设置超参数
loss_fn    = nn.CrossEntropyLoss() # 创建损失函数
learn_rate = 1e-2 # 学习率
opt        = torch.optim.SGD(model.parameters(),lr=learn_rate)
# 训练循环
def train(dataloader, model, loss_fn, optimizer):
    size = len(dataloader.dataset)  # 训练集的大小，一共60000张图片
    num_batches = len(dataloader)   # 批次数目，1875（60000/32）

    train_loss, train_acc = 0, 0  # 初始化训练损失和正确率
    
    for X, y in dataloader:  # 获取图片及其标签
        X, y = X.to(device), y.to(device)
        
        # 计算预测误差
        pred = model(X)          # 网络输出
        loss = loss_fn(pred, y)  # 计算网络输出和真实值之间的差距，targets为真实值，计算二者差值即为损失
        
        # 反向传播
        optimizer.zero_grad()  # grad属性归零
        loss.backward()        # 反向传播
        optimizer.step()       # 每一步自动更新
        
        # 记录acc与loss
        train_acc  += (pred.argmax(1) == y).type(torch.float).sum().item()
        train_loss += loss.item()
            
    train_acc  /= size
    train_loss /= num_batches

    return train_acc, train_loss


def test (dataloader, model, loss_fn):
    size        = len(dataloader.dataset)  # 测试集的大小，一共10000张图片
    num_batches = len(dataloader)          # 批次数目，313（10000/32=312.5，向上取整）
    test_loss, test_acc = 0, 0
    
    # 当不进行训练时，停止梯度更新，节省计算内存消耗
    with torch.no_grad():
        for imgs, target in dataloader:
            imgs, target = imgs.to(device), target.to(device)
            
            # 计算loss
            target_pred = model(imgs)
            loss        = loss_fn(target_pred, target)
            
            test_loss += loss.item()
            test_acc  += (target_pred.argmax(1) == target).type(torch.float).sum().item()

    test_acc  /= size
    test_loss /= num_batches

    return test_acc, test_loss

epochs     = 5
train_loss = []
train_acc  = []
test_loss  = []
test_acc   = []

for epoch in range(epochs):
    model.train()
    epoch_train_acc, epoch_train_loss = train(train_dl, model, loss_fn, opt)
    
    model.eval()
    epoch_test_acc, epoch_test_loss = test(test_dl, model, loss_fn)
    
    train_acc.append(epoch_train_acc)
    train_loss.append(epoch_train_loss)
    test_acc.append(epoch_test_acc)
    test_loss.append(epoch_test_loss)
    
    template = ('Epoch:{:2d}, Train_acc:{:.1f}%, Train_loss:{:.3f}, Test_acc:{:.1f}%，Test_loss:{:.3f}')
    print(template.format(epoch+1, epoch_train_acc*100, epoch_train_loss, epoch_test_acc*100, epoch_test_loss))
print('Done')

Output:

torch.Size([32, 1, 28, 28])
=================================================================
Layer (type:depth-idx) Param #
=================================================================
Model –
├─Conv2d: 1-1 320
├─MaxPool2d: 1-2 –
├─Conv2d: 1-3 18,496
├─MaxPool2d: 1-4 –
├─Linear: 1-5 102,464
├─Linear: 1-6 650
=================================================================
Total params: 121,930
Trainable params: 121,930
Non-trainable params: 0
=================================================================
Epoch: 1, Train_acc:76.6%, Train_loss:0.803, Test_acc:92.8%，Test_loss:0.245
Epoch: 2, Train_acc:94.3%, Train_loss:0.188, Test_acc:96.5%，Test_loss:0.121
Epoch: 3, Train_acc:96.3%, Train_loss:0.120, Test_acc:97.1%，Test_loss:0.095
Epoch: 4, Train_acc:97.2%, Train_loss:0.093, Test_acc:97.7%，Test_loss:0.078
Epoch: 5, Train_acc:97.6%, Train_loss:0.078, Test_acc:98.0%，Test_loss:0.061
Done

Repl Closed

使用自定义损失函数

直接替换交叉熵损失函数即可

loss_fn    = CustomCrossEntropyLoss() # 创建损失函数
learn_rate = 1e-2 # 学习率

class CustomCrossEntropyLoss(torch.nn.Module):
    def __init__(self):
        super(CustomCrossEntropyLoss, self).__init__()

    def forward(self, inputs, targets):
        # 手动实现交叉熵损失函数
        log_softmax = inputs.log_softmax(dim=1)
        loss = -log_softmax[torch.arange(log_softmax.size(0)), targets].mean()
        return loss

Output：

torch.Size([32, 1, 28, 28])
=================================================================
Layer (type:depth-idx) Param #
=================================================================
Model –
├─Conv2d: 1-1 320
├─MaxPool2d: 1-2 –
├─Conv2d: 1-3 18,496
├─MaxPool2d: 1-4 –
├─Linear: 1-5 102,464
├─Linear: 1-6 650
=================================================================
Total params: 121,930
Trainable params: 121,930
Non-trainable params: 0
=================================================================
Epoch: 1, Train_acc:80.1%, Train_loss:0.664, Test_acc:92.9%，Test_loss:0.239
Epoch: 2, Train_acc:94.6%, Train_loss:0.181, Test_acc:96.1%，Test_loss:0.128
Epoch: 3, Train_acc:96.5%, Train_loss:0.117, Test_acc:97.4%，Test_loss:0.082
Epoch: 4, Train_acc:97.2%, Train_loss:0.090, Test_acc:97.7%，Test_loss:0.075
Epoch: 5, Train_acc:97.7%, Train_loss:0.075, Test_acc:98.0%，Test_loss:0.063
Done

Repl Closed

结论

对比官方所给交叉熵损失函数与自定义的损失函数，结果还是有轻微差别的。