P33：完整的模型验证套路（Pytorch小土堆学习笔记）

 训练代码（训练轮数提高到了70轮，学习速度降到了0.005）

完整代码如下：

import torch
import torchvision
from torch import nn
from torch import optim
from torch.nn import Conv2d,MaxPool2d,Flatten,Linear,Sequential
from torch.utils.data import DataLoader
from torch.utils.tensorboard import SummaryWriter
import time

#定义训练的设备
device = torch.device("cuda")
#准备数据集
train_data = torchvision.datasets.CIFAR10(root="./dataset2", train=True, transform=torchvision.transforms.ToTensor(),
                                         download=True)
test_data = torchvision.datasets.CIFAR10(root = "./dataset2",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,batch_size=64)
test_dataloader = DataLoader(test_data,batch_size=64)

#准备训练模型
class qiqi(nn.Module):
    def __init__(self):
        super(qiqi, self).__init__()
        self.model1 = Sequential(
            Conv2d(3, 32, 5, padding=2),    #注意有逗号
            MaxPool2d(2),
            Conv2d(32, 32, 5, padding=2),
            MaxPool2d(2),
            Conv2d(32, 64, 5, padding=2),
            MaxPool2d(2),
            Flatten(),
            Linear(1024, 64),
            Linear(64, 10)
        )
    def forward(self,x):
        x=self.model1(x)
        return x
#创建网络模型
qq=qiqi()
qq = qq.to(device)

#损失函数
loss_fn = nn.CrossEntropyLoss()
loss_fn = loss_fn.to(device)

#优化器
learning_rate = 5e-3
optimizer = torch.optim.SGD(qq.parameters(),lr=learning_rate)

#设置训练网络的一些参数
#记录训练的次数
total_train_step = 0
#记录测试的次数
total_test_step = 0
#训练的轮数
epoch = 70
#记录时间
start_time = time.time()
writer = SummaryWriter("train_gpu2")
for i in range(epoch):
    print("--------第 {} 轮训练开始--------".format(i+1))
    #训练步骤开始
    qq.train() #对现在的网络层没影响，在含有bn层和dropout层的模型中有影响，因为这两个层在训练和测试是不一样的
    for data in train_dataloader:
        imgs,targets = data
        imgs = imgs.to(device)
        targets = targets.to(device)
        outputs = qq(imgs)
        loss = loss_fn(outputs,targets)

        #优化器优化模型
        optimizer.zero_grad()
        loss.backward()
        optimizer.step()

        total_train_step = total_train_step+1
        if total_train_step % 100 == 0:
            end_time = time.time()
            print("本轮用时：{}".format(end_time-start_time) )
            print("训练次数：{}, loss: {}".format(total_train_step,loss.item())) #这里item也可以不用
            writer.add_scalar("train_loss",loss.item(),total_train_step)

    #测试步骤开始
    qq.eval()   #对现在的网络层没影响，在含有bn层和dropout层的模型中有影响，因为这两个层在训练和测试是不一样的
    total_test_loss = 0
    total_accurancy = 0
    with torch.no_grad():
        for data in test_dataloader:
            imgs, targets = data
            imgs = imgs.to(device)
            targets = targets.to(device)
            outputs = qq(imgs)
            loss = loss_fn(outputs, targets)
            total_test_loss = total_test_loss + loss.item()
            accurancy = (outputs.argmax(1) == targets).sum()
            total_accurancy = total_accurancy + accurancy

    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_accurancy/test_data_size))
    writer.add_scalar("test_loss",total_test_loss,total_test_step)
    writer.add_scalar("test_accuracy",total_accurancy/test_data_size,total_test_step)
    total_test_step = total_test_step + 1

    torch.save(qq,"qq_model/qq_{}.pth".format(i))
    print("模型已保存")

writer.close()

 

 

 

 （这里显然是有些过拟合了）拿70轮训练的效果不大好了，下面就拿35轮的训练结果验证吧

 

训练完成之后我们拿其中训练好的“qq_35.pth”这个模型来进行验证（这个模型经过训练准确度已经达到了65%左右）

验证代码如下： 

import torch
import torchvision
from PIL import  Image
from torch import nn
from torch.nn import Conv2d,MaxPool2d,Flatten,Linear,Sequential

#导入图片
image_path = "learn_torch/imgs/dog4.png"
image = Image.open(image_path)
image = image.convert('RGB')#因为png格式是四个通道，除了RGB三通道之外，还有一个透明通道，所以，我们调用此行保留颜色通道
                            #加上这句话之后可以进一步适应不用格式，不同截图软件的图片
print(image)

#图片裁切、格式转化
transform = torchvision.transforms.Compose([torchvision.transforms.Resize((32,32)),
                                            torchvision.transforms.ToTensor()])

image = transform(image)
print(image.shape)

class qiqi(nn.Module):
    def __init__(self):
        super(qiqi, self).__init__()
        self.model1 = Sequential(
            Conv2d(3, 32, 5, padding=2),    #注意有逗号
            MaxPool2d(2),
            Conv2d(32, 32, 5, padding=2),
            MaxPool2d(2),
            Conv2d(32, 64, 5, padding=2),
            MaxPool2d(2),
            Flatten(),
            Linear(1024, 64),
            Linear(64, 10)
        )
    def forward(self,x):
        x=self.model1(x)
        return x

#加载模型
model = torch.load("learn_torch/qq_model/qq_35.pth",map_location=torch.device('cpu')) #因为这个模型是在gpu上训练的，因此需要进行映射
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))

拿小狗的图像为例 

 结果对照一下，是正确的