完整的模型训练套路及GPU的利用

1. train.py

import torch
import torchvision
from torch import nn
from torch.utils.data import DataLoader
from torch.utils.tensorboard import SummaryWriter
import time

#定义训练设备
# device = torch.device('cuda:0')
device = torch.device("cuda"if torch.cuda.is_available() else "cpu")


#准备数据集
train_data = torchvision.datasets.CIFAR10('../dataset',train=True,transform=torchvision.transforms.ToTensor())
test_data = torchvision.datasets.CIFAR10('../dataset',train=False,transform=torchvision.transforms.ToTensor())

#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), '测试数据集的长度为：{}'.format(test_data_size))

#打包数据
train_dataloader = DataLoader(train_data,batch_size=64)
test_dataloader = DataLoader(test_data,batch_size=64)

#创建网络模型  这块也可以单独放一个文件引入
class CIF(nn.Module):
    def __init__(self):
        super(CIF, self).__init__()
        self.model = nn.Sequential(
            nn.Conv2d(3,32,5,1,padding=2),
            nn.MaxPool2d(kernel_size=2),
            nn.Conv2d(32,32,5,1,2),
            nn.MaxPool2d(2),
            nn.Conv2d(32,64,5,1,2),
            nn.MaxPool2d(2),
            nn.Flatten(),
            nn.Linear(1024,64),
            nn.Linear(64,10)
        )

    def forward(self,x):
        x = self.model(x)
        return x

cif = CIF()
cif.to(device)
# cif = cif.cuda()  #网络模型转移到cuda上
#良好的写法：
# if torch.cuda.is_available():
#     cif = cif.cuda()

#损失函数
loss_fn = nn.CrossEntropyLoss()
loss_fn = loss_fn.to(device)
# loss_fn = loss_fn.cuda()  #损失函数转移到cuda

#优化器
 #learing_rate = 0.01 #方便修改
learing_rate = 1e-2  #1x(10)^(-2)
optimzer = torch.optim.SGD(cif.parameters(),lr=learing_rate)


#设置训练网络的一些参数
 #记录训练次数
total_train_step = 0
 #记录测试的次数
total_test_step = 0
 #训练的轮数
epoch = 10

#添加tendorboard
writer = SummaryWriter('../logs')

#计算开始时间
start_time = time.time()

for i in range(epoch):
    print('-----------第{}轮训练开始------------'.format(i+1))

    #训练步骤开始
    cif.train()  #和cif.eval()只对部分网络有用（dropout等）
    for data in train_dataloader:
        imgs, targets = data
        imgs = imgs.to(device)
        targets = targets.to(device)
        # imgs = imgs.cuda()  #图像数据加载到cuda
        # targets = targets.cuda()  #标签加载到cuda
        output = cif(imgs)
        loss = loss_fn(output,targets)

        #优化器调优
        optimzer.zero_grad()
        loss.backward()
        optimzer.step()

        total_train_step = total_train_step+1
        if total_train_step % 100 == 0:   #每100次打印一次
            end_time = time.time()
            print('训练时长:{}'.format(end_time-start_time))   #计算总时长
            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
    cif.eval()
    total_test_loss = 0  #损失指标
    total_accuracy = 0  #准确率指标
    with torch.no_grad():
        for data in test_dataloader:
            imgs, targets = data
            imgs = imgs.to(device)
            targets = targets.to(device)
            # imgs = imgs.cuda()  # 图像数据加载到cuda
            # targets = targets.cuda()  # 标签加载到cuda
            outputs = cif(imgs)
            loss = loss_fn(outputs,targets)
            accuracy = (outputs.argmax(1) == targets).sum()   #1为横着方向
            total_test_loss = total_test_loss+loss.item()
            total_accuracy = total_accuracy+accuracy

    print('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_accuracy/test_data_size))
    writer.add_scalar('test_loss',total_test_loss,total_test_step)
    writer.add_scalar('test_accuracy',total_accuracy/test_data_size,total_test_step)

    total_test_step = total_test_step+1

    #保存模型
    torch.save(cif,'cif_{}.pt'.format(i))
    print('模型已保存')

writer.close()

2. test.py

import torch
import torchvision
from PIL import Image
from torch import nn

img_path = 'dataset/train/dog/dog.jpg'
img = Image.open(img_path)

image = img.convert('RGB')

transforms = torchvision.transforms.Compose([torchvision.transforms.Resize((32,32)),torchvision.transforms.ToTensor()])

image = transforms(image)
# print(image.shape)

class CIF(nn.Module):
    def __init__(self):
        super(CIF, self).__init__()
        self.model = nn.Sequential(
            nn.Conv2d(3,32,5,1,padding=2),
            nn.MaxPool2d(kernel_size=2),
            nn.Conv2d(32,32,5,1,2),
            nn.MaxPool2d(2),
            nn.Conv2d(32,64,5,1,2),
            nn.MaxPool2d(2),
            nn.Flatten(),
            nn.Linear(1024,64),
            nn.Linear(64,10)
        )

    def forward(self,x):
        x = self.model(x)
        return x


model = torch.load('gpu_train/cif_5.pt', map_location=torch.device('cpu')) #GPU训练的模型 cpu时需要映射过来
print(model)
image = torch.reshape(image,(1,3,32,32))
model.eval()
with torch.no_grad():
    output = model(image)
print(output)

print(output.argmax(1))