1.准备数据集

train_data = torchvision.datasets.CIFAR10("../data",train=True,transform=torchvision.transforms.ToTensor(),download=True)
test_data = torchvision.datasets.CIFAR10("../data",train=False,transform=torchvision.transforms.ToTensor(),download=True)

2.使用DataLoader加载数据

#利用DataLoader 来加载数据集
train_dataloader = DataLoader(train_data_size,batch_size=64)
test_dataloader = DataLoader(test_data_size,batch_size=64)

3.创建模型(根据需要调整)

class My_Module(nn.Module):def __init__(self):super(My_Module,self).__init__()          XXXdef forward(self,x):return x

4.加载模型

#创建网络模型
My_Module = My_Module()#损失函数
loss_fn = nn.CrossEntropyLoss()#优化器
learning_rate = 0.0001
optimizer = torch.optim.SGD(My_Module.patameters(),lr=learning_rate)#设置训练网络的参数
total_train_step = 0    #记录训练的次数
total_test_step = 0     #记录测试的次数
epoch = 100     #训练的轮数#添加tensorboard
writer = SummaryWriter("../logs_train")

5.训练

for i in range(epoch):print("-----------第 {} 训练开始-----------".format(i+1))#训练步骤开始My_Module.train()for data in train_dataloader:imgs,labels = dataoutputs = My_Module(imgs)loss = loss_fn(outputs,labels)#优化器优化模型optimizer.zero_grad()loss.backward()optimizer.step()total_train_step = total_train_step + 1if total_train_step % 100 == 0:print("训练次数：{}，loss:{}".format(total_train_step,loss.item()))writer.add_scalar("train_loss", loss.item(),total_train_step)#测试步骤开始My_Module.eval()total_test_loss = 0total_arruracy = 0with torch.no_grad():for data in test_dataloader:test_imgs, test_labels = dataoutputs = My_Module(test_imgs)loss = loss_fn(outputs,test_labels)total_test_loss = total_test_loss +loss.item()accuracy = (outputs.argmax(1) == test_labels).sum()total_arruracy = total_arruracy + accuracyprint("整体测试集上的loss:{}".format(total_test_loss))print("整体测试集上的正确率:{}".format(total_arruracy/test_data_size))writer.add_scalar("test_loss", total_test_loss,total_test_step)writer.add_scalar("test_accuracy", total_arruracy/test_data_size,total_test_step)torch.save(My_Module,"My_Module_{}.pth".format(i))print("模型已保存")writer.close()

6.查看loss图

1.在项目终端输入：tensorboard --logdir=logs
2.在浏览器中打开就可以看到loss图像了
说明：tensorboard --logdir=logs中的“logs”指的是SummaryWriter("logs")中的“logs”。这两个保持一致就可以了。

完整代码

# -*- coding: utf-8 -*-
"""
Created on Fri Apr 12 20:54:46 2024@author: ZW
"""import torchvision
from torch.utils.tensorboard import SummaryWriter
from torch import nn
from torch.utils.data import DataLoader
import torchtrain_data = torchvision.datasets.CIFAR10("../data",train=True,transform=torchvision.transforms.ToTensor(),download=True)
test_data = torchvision.datasets.CIFAR10("../data",train=False,transform=torchvision.transforms.ToTensor(),download=True)#length长度
train_data_size = len(train_data)
test_data_size = len(test_data)
print("训练数据集的长度为：{}".format(train_data_size))
print("测试数据集的长度为：{}".format(test_data_size))#利用DataLoader 来加载数据集
train_dataloader = DataLoader(train_data_size,batch_size=64)
test_dataloader = DataLoader(test_data_size,batch_size=64)class My_Module(nn.Module):def __init__(self):super(My_Module,self).__init__()          def forward(self,x):return x#创建网络模型
My_Module = My_Module()#损失函数
loss_fn = nn.CrossEntropyLoss()#优化器
learning_rate = 0.0001
optimizer = torch.optim.SGD(My_Module.patameters(),lr=learning_rate)#设置训练网络的参数
total_train_step = 0    #记录训练的次数
total_test_step = 0     #记录测试的次数
epoch = 100     #训练的轮数#添加tensorboard
writer = SummaryWriter("../logs_train")for i in range(epoch):print("-----------第 {} 训练开始-----------".format(i+1))#训练步骤开始My_Module.train()for data in train_dataloader:imgs,labels = dataoutputs = My_Module(imgs)loss = loss_fn(outputs,labels)#优化器优化模型optimizer.zero_grad()loss.backward()optimizer.step()total_train_step = total_train_step + 1if total_train_step % 100 == 0:print("训练次数：{}，loss:{}".format(total_train_step,loss.item()))writer.add_scalar("train_loss", loss.item(),total_train_step)#测试步骤开始My_Module.eval()total_test_loss = 0total_arruracy = 0with torch.no_grad():for data in test_dataloader:test_imgs, test_labels = dataoutputs = My_Module(test_imgs)loss = loss_fn(outputs,test_labels)total_test_loss = total_test_loss +loss.item()accuracy = (outputs.argmax(1) == test_labels).sum()total_arruracy = total_arruracy + accuracyprint("整体测试集上的loss:{}".format(total_test_loss))print("整体测试集上的正确率:{}".format(total_arruracy/test_data_size))writer.add_scalar("test_loss", total_test_loss,total_test_step)writer.add_scalar("test_accuracy", total_arruracy/test_data_size,total_test_step)torch.save(My_Module,"My_Module_{}.pth".format(i))print("模型已保存")writer.close()