深度学习Day-07：实现咖啡豆识别（VGG-16复现）

Point__Nemo

已于 2024-03-01 17:09:26 修改

阅读量813

点赞数 22

分类专栏：深度学习文章标签：深度学习人工智能

于 2024-02-26 16:15:51 首次发布

本文链接：https://blog.csdn.net/m0_51359915/article/details/136300587

版权

深度学习专栏收录该内容

8 篇文章 0 订阅

订阅专栏

🍨 本文为：[🔗365天深度学习训练营] 中的学习记录博客
🍖 原作者：[K同学啊 | 接辅导、项目定制]

一、基础配置

语言环境：Python3.8
编译器选择：Pycharm
深度学习环境：

- torch==1.12.1+cu113
- torchvision==0.13.1+cu113

二、前期准备

1.设置GPU

import torch
import torch.nn as nn
from torchvision import transforms, datasets
import warnings

warnings.filterwarnings("ignore")       #忽略警告信息

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
print(device)

根据个人设备情况，选择使用GPU/CPU进行训练，在Pycharm中需要添加print命令来查看是否使用了GPU ，若GPU可用则输出

cuda

该代码片段中加入了 warnings.filterwarnings("ignore")用于忽略代码运行中不必要的警告信息。

2. 导入数据

本项目所采用的数据集未收录于公开数据中，故需要自己在文件目录中导入相应数据集合，并设置对应文件目录，以供后续学习过程中使用。

运行下述代码：

import pathlib

data_dir = 'data'
data_dir = pathlib.Path(data_dir)

data_paths  = list(data_dir.glob('*'))
classeNames = [str(path).split("\\")[1] for path in data_paths]
print(classeNames)

得到如下输出：

['Dark', 'Green', 'Light', 'Medium']

接下来，我们通过transforms.Compose对整个数据集进行预处理：

第一步：将输入图片resize成统一尺寸，即[224, 224]
第二步：转换为tensor，并归一化到[0,1]之间
第三步：转换为标准正态分布（高斯分布），使模型更容易收敛

train_transforms = transforms.Compose([
    transforms.Resize([224, 224]),      # 将输入图片resize成统一尺寸
    transforms.RandomHorizontalFlip(),  # 随机水平翻转
    transforms.ToTensor(),              # 将PIL Image或numpy.ndarray转换为tensor，并归一化到[0,1]之间
    transforms.Normalize(               # 标准化处理-->转换为标准正太分布（高斯分布），使模型更容易收敛
        mean=[0.485, 0.456, 0.406],
        std=[0.229, 0.224, 0.225])      # 其中 mean=[0.485,0.456,0.406]与std=[0.229,0.224,0.225] 从数据集中随机抽样计算得到的。
])

test_transform = transforms.Compose([
    transforms.Resize([224, 224]),  # 将输入图片resize成统一尺寸
    transforms.ToTensor(),          # 将PIL Image或numpy.ndarray转换为tensor，并归一化到[0,1]之间
    transforms.Normalize(           # 标准化处理-->转换为标准正太分布（高斯分布），使模型更容易收敛
        mean=[0.485, 0.456, 0.406],
        std=[0.229, 0.224, 0.225])  # 其中 mean=[0.485,0.456,0.406]与std=[0.229,0.224,0.225] 从数据集中随机抽样计算得到的。
])

total_data = datasets.ImageFolder("data",transform=train_transforms)
print(total_data)

得到如下输出：

Dataset ImageFolder
    Number of datapoints: 1200
    Root location: data
    StandardTransform
Transform: Compose(
               Resize(size=[224, 224], interpolation=bilinear, max_size=None, antialias=None)
               RandomHorizontalFlip(p=0.5)
               ToTensor()
               Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
           )

接下来，为了方便模型进行训练和推理，映射关系可以将类别标签转换为模型可以理解的数字格式：

total_data.class_to_idx

得到如下输出：

{'Dark': 0, 'Green': 1, 'Light': 2, 'Medium': 3}

3. 划分数据集

此处数据集需要做按比例划分的操作：

train_size = int(0.8 * len(total_data))
test_size  = len(total_data) - train_size
train_dataset, test_dataset = torch.utils.data.random_split(total_data, [train_size, test_size])
print(train_dataset)
print(test_dataset)

根据代码所示，本文将原数据集按照8：2的比例进行了划分及打印：

<torch.utils.data.dataset.Subset object at 0x000001ACDB4237F0>
<torch.utils.data.dataset.Subset object at 0x000001ACDB42E730>

接下来，根据划分得到的训练集和验证集对数据集进行包装：

batch_size = 32

train_dl = torch.utils.data.DataLoader(train_dataset,
                                           batch_size=batch_size,
                                           shuffle=True,
                                           num_workers=0)
test_dl = torch.utils.data.DataLoader(test_dataset,
                                          batch_size=batch_size,
                                          shuffle=True,
                                          num_workers=0)

并通过：

for X, y in test_dl:
    print("Shape of X [N, C, H, W]: ", X.shape)
    print("Shape of y: ", y.shape, y.dtype)
    break

输出测试数据集的数据分布情况：

Shape of X [N, C, H, W]:  torch.Size([32, 3, 224, 224])
Shape of y:  torch.Size([32]) torch.int64

Tips：win需要采用num_workers=0的模式。

4.手动搭建VGG-16模型

1.模型搭建

VGG-16结构说明：

13个卷积层（Convolutional Layer），分别用blockX_convX表示；
3个全连接层（Fully connected Layer），用classifier表示；
5个池化层（Pool layer）；

VGG-16包含了16个隐藏层（13个卷积层和3个全连接层），故称为VGG-16

class vgg16(nn.Module):
    def __init__(self):
        super(vgg16, self).__init__()
        # 卷积块1
        self.block1 = nn.Sequential(
            nn.Conv2d(3, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)),
            nn.ReLU(),
            nn.Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=(2, 2), stride=(2, 2))
        )
        # 卷积块2
        self.block2 = nn.Sequential(
            nn.Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)),
            nn.ReLU(),
            nn.Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=(2, 2), stride=(2, 2))
        )
        # 卷积块3
        self.block3 = nn.Sequential(
            nn.Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)),
            nn.ReLU(),
            nn.Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)),
            nn.ReLU(),
            nn.Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=(2, 2), stride=(2, 2))
        )
        # 卷积块4
        self.block4 = nn.Sequential(
            nn.Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)),
            nn.ReLU(),
            nn.Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)),
            nn.ReLU(),
            nn.Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=(2, 2), stride=(2, 2))
        )
        # 卷积块5
        self.block5 = nn.Sequential(
            nn.Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)),
            nn.ReLU(),
            nn.Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)),
            nn.ReLU(),
            nn.Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=(2, 2), stride=(2, 2))
        )

        # 全连接网络层，用于分类
        self.classifier = nn.Sequential(
            nn.Linear(in_features=512 * 7 * 7, out_features=4096),
            nn.ReLU(),
            nn.Linear(in_features=4096, out_features=4096),
            nn.ReLU(),
            nn.Linear(in_features=4096, out_features=4)
        )

    def forward(self, x):
        x = self.block1(x)
        x = self.block2(x)
        x = self.block3(x)
        x = self.block4(x)
        x = self.block5(x)
        x = torch.flatten(x, start_dim=1)
        x = self.classifier(x)

        return x


device = "cuda" if torch.cuda.is_available() else "cpu"
print("Using {} device".format(device))

model = vgg16().to(device)
print(model)

可以得到如下输出：

Using cuda device
vgg16(
  (block1): Sequential(
    (0): Conv2d(3, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (1): ReLU()
    (2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (3): ReLU()
    (4): MaxPool2d(kernel_size=(2, 2), stride=(2, 2), padding=0, dilation=1, ceil_mode=False)
  )
  (block2): Sequential(
    (0): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (1): ReLU()
    (2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (3): ReLU()
    (4): MaxPool2d(kernel_size=(2, 2), stride=(2, 2), padding=0, dilation=1, ceil_mode=False)
  )
  (block3): Sequential(
    (0): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (1): ReLU()
    (2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (3): ReLU()
    (4): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (5): ReLU()
    (6): MaxPool2d(kernel_size=(2, 2), stride=(2, 2), padding=0, dilation=1, ceil_mode=False)
  )
  (block4): Sequential(
    (0): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (1): ReLU()
    (2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (3): ReLU()
    (4): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (5): ReLU()
    (6): MaxPool2d(kernel_size=(2, 2), stride=(2, 2), padding=0, dilation=1, ceil_mode=False)
  )
  (block5): Sequential(
    (0): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (1): ReLU()
    (2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (3): ReLU()
    (4): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (5): ReLU()
    (6): MaxPool2d(kernel_size=(2, 2), stride=(2, 2), padding=0, dilation=1, ceil_mode=False)
  )
  (classifier): Sequential(
    (0): Linear(in_features=25088, out_features=4096, bias=True)
    (1): ReLU()
    (2): Linear(in_features=4096, out_features=4096, bias=True)
    (3): ReLU()
    (4): Linear(in_features=4096, out_features=4, bias=True)
  )
)

2.查看模型信息

# 统计模型参数量以及其他指标
import torchsummary as summary
summary.summary(model, (3, 224, 224))

得到如下输出：

----------------------------------------------------------------
        Layer (type)               Output Shape         Param #
================================================================
            Conv2d-1         [-1, 64, 224, 224]           1,792
              ReLU-2         [-1, 64, 224, 224]               0
            Conv2d-3         [-1, 64, 224, 224]          36,928
              ReLU-4         [-1, 64, 224, 224]               0
         MaxPool2d-5         [-1, 64, 112, 112]               0
            Conv2d-6        [-1, 128, 112, 112]          73,856
              ReLU-7        [-1, 128, 112, 112]               0
            Conv2d-8        [-1, 128, 112, 112]         147,584
              ReLU-9        [-1, 128, 112, 112]               0
        MaxPool2d-10          [-1, 128, 56, 56]               0
           Conv2d-11          [-1, 256, 56, 56]         295,168
             ReLU-12          [-1, 256, 56, 56]               0
           Conv2d-13          [-1, 256, 56, 56]         590,080
             ReLU-14          [-1, 256, 56, 56]               0
           Conv2d-15          [-1, 256, 56, 56]         590,080
             ReLU-16          [-1, 256, 56, 56]               0
        MaxPool2d-17          [-1, 256, 28, 28]               0
           Conv2d-18          [-1, 512, 28, 28]       1,180,160
             ReLU-19          [-1, 512, 28, 28]               0
           Conv2d-20          [-1, 512, 28, 28]       2,359,808
             ReLU-21          [-1, 512, 28, 28]               0
           Conv2d-22          [-1, 512, 28, 28]       2,359,808
             ReLU-23          [-1, 512, 28, 28]               0
        MaxPool2d-24          [-1, 512, 14, 14]               0
           Conv2d-25          [-1, 512, 14, 14]       2,359,808
             ReLU-26          [-1, 512, 14, 14]               0
           Conv2d-27          [-1, 512, 14, 14]       2,359,808
             ReLU-28          [-1, 512, 14, 14]               0
           Conv2d-29          [-1, 512, 14, 14]       2,359,808
             ReLU-30          [-1, 512, 14, 14]               0
        MaxPool2d-31            [-1, 512, 7, 7]               0
           Linear-32                 [-1, 4096]     102,764,544
             ReLU-33                 [-1, 4096]               0
           Linear-34                 [-1, 4096]      16,781,312
             ReLU-35                 [-1, 4096]               0
           Linear-36                    [-1, 4]          16,388
================================================================
Total params: 134,276,932
Trainable params: 134,276,932
Non-trainable params: 0
----------------------------------------------------------------
Input size (MB): 0.57
Forward/backward pass size (MB): 218.52
Params size (MB): 512.23
Estimated Total Size (MB): 731.32
----------------------------------------------------------------

参数量为：134,276,932

三、训练模型

1. 编写训练函数

# 训练循环
def train(dataloader, model, loss_fn, optimizer):
    size = len(dataloader.dataset)  # 训练集的大小
    num_batches = len(dataloader)   # 批次数目, (size/batch_size，向上取整)

    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

2. 编写测试函数

测试函数和训练函数大致相同，但是由于不进行梯度下降对网络权重进行更新，所以不需要传入优化器

def test(dataloader, model, loss_fn):
    size = len(dataloader.dataset)  # 测试集的大小
    num_batches = len(dataloader)   # 批次数目, (size/batch_size，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

3.正式训练

import copy

optimizer = torch.optim.Adam(model.parameters(), lr=1e-4)
loss_fn = nn.CrossEntropyLoss()  # 创建损失函数

epochs = 40

train_loss = []
train_acc = []
test_loss = []
test_acc = []

best_acc = 0  # 设置一个最佳准确率，作为最佳模型的判别指标

for epoch in range(epochs):

    model.train()
    epoch_train_acc, epoch_train_loss = train(train_dl, model, loss_fn, optimizer)

    model.eval()
    epoch_test_acc, epoch_test_loss = test(test_dl, model, loss_fn)

    # 保存最佳模型到 best_model
    if epoch_test_acc > best_acc:
        best_acc = epoch_test_acc
        best_model = copy.deepcopy(model)

    train_acc.append(epoch_train_acc)
    train_loss.append(epoch_train_loss)
    test_acc.append(epoch_test_acc)
    test_loss.append(epoch_test_loss)

    # 获取当前的学习率
    lr = optimizer.state_dict()['param_groups'][0]['lr']

    template = ('Epoch:{:2d}, Train_acc:{:.1f}%, Train_loss:{:.3f}, Test_acc:{:.1f}%, Test_loss:{:.3f}, Lr:{:.2E}')
    print(template.format(epoch + 1, epoch_train_acc * 100, epoch_train_loss,
                          epoch_test_acc * 100, epoch_test_loss, lr))

# 保存最佳模型到文件中
PATH = './best_model.pth'  # 保存的参数文件名
torch.save(model.state_dict(), PATH)

print('Done')

得到如下输出：

Epoch: 1, Train_acc:25.9%, Train_loss:1.387, Test_acc:27.1%, Test_loss:1.339, Lr:1.00E-04
Epoch: 2, Train_acc:48.2%, Train_loss:1.009, Test_acc:62.5%, Test_loss:0.872, Lr:1.00E-04
Epoch: 3, Train_acc:65.6%, Train_loss:0.708, Test_acc:72.9%, Test_loss:0.673, Lr:1.00E-04
Epoch: 4, Train_acc:68.2%, Train_loss:0.668, Test_acc:73.3%, Test_loss:0.558, Lr:1.00E-04
Epoch: 5, Train_acc:75.3%, Train_loss:0.517, Test_acc:80.0%, Test_loss:0.436, Lr:1.00E-04
Epoch: 6, Train_acc:75.6%, Train_loss:0.489, Test_acc:84.6%, Test_loss:0.360, Lr:1.00E-04
Epoch: 7, Train_acc:79.0%, Train_loss:0.467, Test_acc:75.8%, Test_loss:0.432, Lr:1.00E-04
Epoch: 8, Train_acc:81.9%, Train_loss:0.409, Test_acc:83.8%, Test_loss:0.299, Lr:1.00E-04
Epoch: 9, Train_acc:88.6%, Train_loss:0.263, Test_acc:91.2%, Test_loss:0.203, Lr:1.00E-04
Epoch:10, Train_acc:95.1%, Train_loss:0.126, Test_acc:93.3%, Test_loss:0.214, Lr:1.00E-04
Epoch:11, Train_acc:94.7%, Train_loss:0.140, Test_acc:92.1%, Test_loss:0.174, Lr:1.00E-04
Epoch:12, Train_acc:96.9%, Train_loss:0.098, Test_acc:98.8%, Test_loss:0.027, Lr:1.00E-04
Epoch:13, Train_acc:98.4%, Train_loss:0.057, Test_acc:98.8%, Test_loss:0.026, Lr:1.00E-04
Epoch:14, Train_acc:96.9%, Train_loss:0.105, Test_acc:98.8%, Test_loss:0.056, Lr:1.00E-04
Epoch:15, Train_acc:97.8%, Train_loss:0.066, Test_acc:99.2%, Test_loss:0.050, Lr:1.00E-04
Epoch:16, Train_acc:94.4%, Train_loss:0.171, Test_acc:92.9%, Test_loss:0.163, Lr:1.00E-04
Epoch:17, Train_acc:95.8%, Train_loss:0.126, Test_acc:93.3%, Test_loss:0.166, Lr:1.00E-04
Epoch:18, Train_acc:95.9%, Train_loss:0.113, Test_acc:95.8%, Test_loss:0.092, Lr:1.00E-04
Epoch:19, Train_acc:97.1%, Train_loss:0.082, Test_acc:98.8%, Test_loss:0.040, Lr:1.00E-04
Epoch:20, Train_acc:99.2%, Train_loss:0.032, Test_acc:100.0%, Test_loss:0.009, Lr:1.00E-04
Epoch:21, Train_acc:98.8%, Train_loss:0.032, Test_acc:93.8%, Test_loss:0.145, Lr:1.00E-04
Epoch:22, Train_acc:98.9%, Train_loss:0.036, Test_acc:100.0%, Test_loss:0.006, Lr:1.00E-04
Epoch:23, Train_acc:99.8%, Train_loss:0.010, Test_acc:96.2%, Test_loss:0.108, Lr:1.00E-04
Epoch:24, Train_acc:97.9%, Train_loss:0.051, Test_acc:97.5%, Test_loss:0.060, Lr:1.00E-04
Epoch:25, Train_acc:96.4%, Train_loss:0.122, Test_acc:99.6%, Test_loss:0.030, Lr:1.00E-04
Epoch:26, Train_acc:98.9%, Train_loss:0.035, Test_acc:99.2%, Test_loss:0.019, Lr:1.00E-04
Epoch:27, Train_acc:98.8%, Train_loss:0.038, Test_acc:97.9%, Test_loss:0.049, Lr:1.00E-04
Epoch:28, Train_acc:99.1%, Train_loss:0.030, Test_acc:98.8%, Test_loss:0.038, Lr:1.00E-04
Epoch:29, Train_acc:99.3%, Train_loss:0.016, Test_acc:99.6%, Test_loss:0.014, Lr:1.00E-04
Epoch:30, Train_acc:99.9%, Train_loss:0.007, Test_acc:99.6%, Test_loss:0.005, Lr:1.00E-04
Epoch:31, Train_acc:99.7%, Train_loss:0.005, Test_acc:96.2%, Test_loss:0.109, Lr:1.00E-04
Epoch:32, Train_acc:97.6%, Train_loss:0.074, Test_acc:98.8%, Test_loss:0.043, Lr:1.00E-04
Epoch:33, Train_acc:98.1%, Train_loss:0.045, Test_acc:99.2%, Test_loss:0.032, Lr:1.00E-04
Epoch:34, Train_acc:99.1%, Train_loss:0.036, Test_acc:98.3%, Test_loss:0.047, Lr:1.00E-04
Epoch:35, Train_acc:99.0%, Train_loss:0.026, Test_acc:98.8%, Test_loss:0.080, Lr:1.00E-04
Epoch:36, Train_acc:98.8%, Train_loss:0.025, Test_acc:99.6%, Test_loss:0.074, Lr:1.00E-04
Epoch:37, Train_acc:99.3%, Train_loss:0.011, Test_acc:100.0%, Test_loss:0.001, Lr:1.00E-04
Epoch:38, Train_acc:99.2%, Train_loss:0.029, Test_acc:99.6%, Test_loss:0.012, Lr:1.00E-04
Epoch:39, Train_acc:98.4%, Train_loss:0.034, Test_acc:97.9%, Test_loss:0.086, Lr:1.00E-04
Epoch:40, Train_acc:99.4%, Train_loss:0.019, Test_acc:99.2%, Test_loss:0.014, Lr:1.00E-04
Done

四、结果可视化

1. Loss&Accuracy

import matplotlib.pyplot as plt
#隐藏警告
import warnings
warnings.filterwarnings("ignore")               #忽略警告信息
plt.rcParams['font.sans-serif']    = ['SimHei'] # 用来正常显示中文标签
plt.rcParams['axes.unicode_minus'] = False      # 用来正常显示负号
plt.rcParams['figure.dpi']         = 100        #分辨率

epochs_range = range(epochs)

plt.figure(figsize=(12, 3))
plt.subplot(1, 2, 1)

plt.plot(epochs_range, train_acc, label='Training Accuracy')
plt.plot(epochs_range, test_acc, label='Test Accuracy')
plt.legend(loc='lower right')
plt.title('Training and Validation Accuracy')

plt.subplot(1, 2, 2)
plt.plot(epochs_range, train_loss, label='Training Loss')
plt.plot(epochs_range, test_loss, label='Test Loss')
plt.legend(loc='upper right')
plt.title('Training and Validation Loss')
plt.show()

得到的可视化结果：

2. 指定图片进行预测

首先，先定义出一个用于预测的函数：

from PIL import Image

classes = list(total_data.class_to_idx)

def predict_one_image(image_path, model, transform, classes):
    test_img = Image.open(image_path).convert('RGB')
    plt.imshow(test_img)  # 展示预测的图片

    test_img = transform(test_img)
    img = test_img.to(device).unsqueeze(0)

    model.eval()
    output = model(img)

    _, pred = torch.max(output, 1)
    pred_class = classes[pred]
    print(f'预测结果是：{pred_class}')

接着调用函数对指定图片进行预测：

# 预测训练集中的某张照片
predict_one_image(image_path='data/Dark/dark (1).png',
                  model=model,
                  transform=train_transforms,
                  classes=classes)

得到如下结果：

预测结果是：Dark

3.模型评估

将模型调至评估模式：

best_model.eval()
epoch_test_acc, epoch_test_loss = test(test_dl, best_model, loss_fn)
print(epoch_test_acc)
print(epoch_test_loss)
print(epoch_test_acc)

得到如下输出：

1.0
0.010690234747016802
1.0

观察得到和前文中一致。

五、个人理解

本项目需要根据所给的图片，来识别咖啡豆情况。

根据项目基本要求：

自己搭建VGG-16网络框架；
调用官方的VGG-16网络框架；
查看模型的参数量以及相关指标；

目前均已完成，本文的主要更新为自己搭建VGG-16实现咖啡豆识别。

特别的，针对要求②所提出的调用原生VGG的方法，可参考Day-06中模型构建的方法，需要指出的是，调用原生VGG时，所展现出的网络模型为：

Using cuda device
VGG(
  (features): Sequential(
    (0): Conv2d(3, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (1): ReLU(inplace=True)
    (2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (3): ReLU(inplace=True)
    (4): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (5): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (6): ReLU(inplace=True)
    (7): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (8): ReLU(inplace=True)
    (9): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (10): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (11): ReLU(inplace=True)
    (12): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (13): ReLU(inplace=True)
    (14): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (15): ReLU(inplace=True)
    (16): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (17): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (18): ReLU(inplace=True)
    (19): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (20): ReLU(inplace=True)
    (21): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (22): ReLU(inplace=True)
    (23): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (24): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (25): ReLU(inplace=True)
    (26): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (27): ReLU(inplace=True)
    (28): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (29): ReLU(inplace=True)
    (30): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  )
  (avgpool): AdaptiveAvgPool2d(output_size=(7, 7))
  (classifier): Sequential(
    (0): Linear(in_features=25088, out_features=4096, bias=True)
    (1): ReLU(inplace=True)
    (2): Dropout(p=0.5, inplace=False)
    (3): Linear(in_features=4096, out_features=4096, bias=True)
    (4): ReLU(inplace=True)
    (5): Dropout(p=0.5, inplace=False)
    (6): Linear(in_features=4096, out_features=4, bias=True)
  )
)

参数量为：

----------------------------------------------------------------
        Layer (type)               Output Shape         Param #
================================================================
            Conv2d-1         [-1, 64, 224, 224]           1,792
              ReLU-2         [-1, 64, 224, 224]               0
            Conv2d-3         [-1, 64, 224, 224]          36,928
              ReLU-4         [-1, 64, 224, 224]               0
         MaxPool2d-5         [-1, 64, 112, 112]               0
            Conv2d-6        [-1, 128, 112, 112]          73,856
              ReLU-7        [-1, 128, 112, 112]               0
            Conv2d-8        [-1, 128, 112, 112]         147,584
              ReLU-9        [-1, 128, 112, 112]               0
        MaxPool2d-10          [-1, 128, 56, 56]               0
           Conv2d-11          [-1, 256, 56, 56]         295,168
             ReLU-12          [-1, 256, 56, 56]               0
           Conv2d-13          [-1, 256, 56, 56]         590,080
             ReLU-14          [-1, 256, 56, 56]               0
           Conv2d-15          [-1, 256, 56, 56]         590,080
             ReLU-16          [-1, 256, 56, 56]               0
        MaxPool2d-17          [-1, 256, 28, 28]               0
           Conv2d-18          [-1, 512, 28, 28]       1,180,160
             ReLU-19          [-1, 512, 28, 28]               0
           Conv2d-20          [-1, 512, 28, 28]       2,359,808
             ReLU-21          [-1, 512, 28, 28]               0
           Conv2d-22          [-1, 512, 28, 28]       2,359,808
             ReLU-23          [-1, 512, 28, 28]               0
        MaxPool2d-24          [-1, 512, 14, 14]               0
           Conv2d-25          [-1, 512, 14, 14]       2,359,808
             ReLU-26          [-1, 512, 14, 14]               0
           Conv2d-27          [-1, 512, 14, 14]       2,359,808
             ReLU-28          [-1, 512, 14, 14]               0
           Conv2d-29          [-1, 512, 14, 14]       2,359,808
             ReLU-30          [-1, 512, 14, 14]               0
        MaxPool2d-31            [-1, 512, 7, 7]               0
AdaptiveAvgPool2d-32            [-1, 512, 7, 7]               0
           Linear-33                 [-1, 4096]     102,764,544
             ReLU-34                 [-1, 4096]               0
          Dropout-35                 [-1, 4096]               0
           Linear-36                 [-1, 4096]      16,781,312
             ReLU-37                 [-1, 4096]               0
          Dropout-38                 [-1, 4096]               0
           Linear-39                    [-1, 4]          16,388
================================================================
Total params: 134,276,932
Trainable params: 16,388
Non-trainable params: 134,260,544
----------------------------------------------------------------
Input size (MB): 0.57
Forward/backward pass size (MB): 218.77
Params size (MB): 512.23
Estimated Total Size (MB): 731.57
----------------------------------------------------------------

训练结果为：

Epoch: 1, Train_acc:38.1%, Train_loss:1.318, Test_acc:78.3%, Test_loss:1.045, Lr:1.00E-04
Epoch: 2, Train_acc:67.0%, Train_loss:0.946, Test_acc:88.3%, Test_loss:0.791, Lr:1.00E-04
Epoch: 3, Train_acc:78.1%, Train_loss:0.770, Test_acc:90.0%, Test_loss:0.656, Lr:1.00E-04
Epoch: 4, Train_acc:84.3%, Train_loss:0.631, Test_acc:90.4%, Test_loss:0.546, Lr:1.00E-04
Epoch: 5, Train_acc:86.7%, Train_loss:0.535, Test_acc:91.7%, Test_loss:0.495, Lr:1.00E-04
Epoch: 6, Train_acc:87.2%, Train_loss:0.478, Test_acc:91.7%, Test_loss:0.430, Lr:1.00E-04
Epoch: 7, Train_acc:89.2%, Train_loss:0.423, Test_acc:93.8%, Test_loss:0.380, Lr:1.00E-04
Epoch: 8, Train_acc:89.9%, Train_loss:0.406, Test_acc:91.7%, Test_loss:0.377, Lr:1.00E-04
Epoch: 9, Train_acc:90.2%, Train_loss:0.368, Test_acc:91.7%, Test_loss:0.341, Lr:1.00E-04
Epoch:10, Train_acc:91.6%, Train_loss:0.345, Test_acc:95.0%, Test_loss:0.318, Lr:1.00E-04
Epoch:11, Train_acc:90.9%, Train_loss:0.337, Test_acc:94.6%, Test_loss:0.299, Lr:1.00E-04
Epoch:12, Train_acc:92.0%, Train_loss:0.329, Test_acc:94.6%, Test_loss:0.294, Lr:1.00E-04
Epoch:13, Train_acc:91.4%, Train_loss:0.309, Test_acc:94.6%, Test_loss:0.277, Lr:1.00E-04
Epoch:14, Train_acc:92.8%, Train_loss:0.283, Test_acc:92.5%, Test_loss:0.271, Lr:1.00E-04
Epoch:15, Train_acc:92.7%, Train_loss:0.279, Test_acc:96.2%, Test_loss:0.244, Lr:1.00E-04
Epoch:16, Train_acc:94.5%, Train_loss:0.246, Test_acc:95.0%, Test_loss:0.235, Lr:1.00E-04
Epoch:17, Train_acc:92.8%, Train_loss:0.263, Test_acc:94.2%, Test_loss:0.227, Lr:1.00E-04
Epoch:18, Train_acc:95.0%, Train_loss:0.237, Test_acc:94.2%, Test_loss:0.245, Lr:1.00E-04
Epoch:19, Train_acc:93.2%, Train_loss:0.246, Test_acc:96.2%, Test_loss:0.225, Lr:1.00E-04
Epoch:20, Train_acc:94.5%, Train_loss:0.231, Test_acc:94.2%, Test_loss:0.232, Lr:1.00E-04
Epoch:21, Train_acc:94.2%, Train_loss:0.228, Test_acc:94.6%, Test_loss:0.217, Lr:1.00E-04
Epoch:22, Train_acc:94.2%, Train_loss:0.227, Test_acc:93.3%, Test_loss:0.219, Lr:1.00E-04
Epoch:23, Train_acc:94.1%, Train_loss:0.218, Test_acc:95.0%, Test_loss:0.212, Lr:1.00E-04
Epoch:24, Train_acc:93.9%, Train_loss:0.223, Test_acc:95.0%, Test_loss:0.191, Lr:1.00E-04
Epoch:25, Train_acc:93.5%, Train_loss:0.215, Test_acc:95.8%, Test_loss:0.199, Lr:1.00E-04
Epoch:26, Train_acc:96.0%, Train_loss:0.185, Test_acc:96.2%, Test_loss:0.197, Lr:1.00E-04
Epoch:27, Train_acc:94.3%, Train_loss:0.203, Test_acc:94.6%, Test_loss:0.196, Lr:1.00E-04
Epoch:28, Train_acc:94.9%, Train_loss:0.193, Test_acc:97.1%, Test_loss:0.175, Lr:1.00E-04
Epoch:29, Train_acc:95.9%, Train_loss:0.183, Test_acc:95.4%, Test_loss:0.186, Lr:1.00E-04
Epoch:30, Train_acc:95.2%, Train_loss:0.178, Test_acc:94.6%, Test_loss:0.179, Lr:1.00E-04
Epoch:31, Train_acc:94.2%, Train_loss:0.197, Test_acc:95.0%, Test_loss:0.206, Lr:1.00E-04
Epoch:32, Train_acc:95.3%, Train_loss:0.176, Test_acc:94.2%, Test_loss:0.183, Lr:1.00E-04
Epoch:33, Train_acc:94.0%, Train_loss:0.192, Test_acc:93.3%, Test_loss:0.193, Lr:1.00E-04
Epoch:34, Train_acc:94.9%, Train_loss:0.181, Test_acc:95.4%, Test_loss:0.172, Lr:1.00E-04
Epoch:35, Train_acc:94.9%, Train_loss:0.168, Test_acc:95.4%, Test_loss:0.165, Lr:1.00E-04
Epoch:36, Train_acc:95.3%, Train_loss:0.171, Test_acc:95.4%, Test_loss:0.168, Lr:1.00E-04
Epoch:37, Train_acc:95.9%, Train_loss:0.159, Test_acc:95.8%, Test_loss:0.160, Lr:1.00E-04
Epoch:38, Train_acc:95.9%, Train_loss:0.163, Test_acc:95.0%, Test_loss:0.164, Lr:1.00E-04
Epoch:39, Train_acc:95.0%, Train_loss:0.168, Test_acc:96.2%, Test_loss:0.154, Lr:1.00E-04
Epoch:40, Train_acc:95.0%, Train_loss:0.156, Test_acc:95.8%, Test_loss:0.158, Lr:1.00E-04

可以看出，调用原生VGG和手动搭建VGG参数量及模型基本一致，这里的准确下降的原因猜测是加入了Dropout，使得整体准确率下降，具体原因有待深入考证。

根据项目拔高要求：