【打卡】006 P6 VGG-16算法-Pytorch实现人脸识别
·
>- **🍨 本文为[🔗365天深度学习训练营](https://mp.weixin.qq.com/s/0dvHCaOoFnW8SCp3JpzKxg) 中的学习记录博客**
>- **🍖 原作者：[K同学啊](https://mtyjkh.blog.csdn.net/)**
打卡～
555环境：
- 语言环境：Python
- 编译器：Jupyter Lab
- 深度学习环境：Pytorch
-
- torch
-
- torchvision
# 一、前期准备
# 1.1 设置环境
import torch
import torch.nn as nn
import torchvision.transforms as transforms
import torchvision
from torchvision import transforms, datasets
import os,PIL,pathlib,warnings
warnings.filterwarnings("ignore") #忽略警告信息
device = torch.device("cpu")
device
# 1.2 导入数据
import os,PIL,random,pathlib
data_dir = '/Users/corawoo/Desktop/data6'
data_dir = pathlib.Path(data_dir)
data_paths = list(data_dir.glob('*'))
classeNames = []
for path in data_paths:
try:
classeName = str(path).split("\\")[1]
classeNames.append(classeName)
except IndexError:
# 处理索引超出范围的情况，可以选择跳过当前路径或进行其他处理
pass
classeNames
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。
])
total_data = datasets.ImageFolder("/Users/corawoo/Desktop/data6",transform=train_transforms)
total_data
Dataset ImageFolder
Number of datapoints: 1800
Root location: /Users/corawoo/Desktop/data6
StandardTransform
Transform: Compose(
Resize(size=[224, 224], interpolation=bilinear, max_size=None, antialias=True)
ToTensor()
Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
)
total_data.class_to_idx
{'Angelina Jolie': 0,
'Brad Pitt': 1,
'Denzel Washington': 2,
'Hugh Jackman': 3,
'Jennifer Lawrence': 4,
'Johnny Depp': 5,
'Kate Winslet': 6,
'Leonardo DiCaprio': 7,
'Megan Fox': 8,
'Natalie Portman': 9,
'Nicole Kidman': 10,
'Robert Downey Jr': 11,
'Sandra Bullock': 12,
'Scarlett Johansson': 13,
'Tom Cruise': 14,
'Tom Hanks': 15,
'Will Smith': 16}
# 1.3 划分数据集
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])
train_dataset, test_dataset
(<torch.utils.data.dataset.Subset at 0x159542f10>, <torch.utils.data.dataset.Subset at 0x1595d3c40>)
batch_size = 32
train_dl = torch.utils.data.DataLoader(train_dataset,
batch_size=batch_size,
shuffle=True,
num_workers=1)
test_dl = torch.utils.data.DataLoader(test_dataset,
batch_size=batch_size,
shuffle=True,
num_workers=1)
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
Shape of X [N, C, H, W]: torch.Size([32, 3, 224, 224]) Shape of y: torch.Size([32]) torch.int64
# 二、调用官方的vgg16模型
VGG-16（Visual Geometry Group-16）是由牛津大学视觉几何组（Visual Geometry Group）提出的一种深度卷积神经网络架构，用于图像分类和对象识别任务。VGG-16在2014年被提出，是VGG系列中的一种。VGG-16之所以备受关注，是因为它在ImageNet图像识别竞赛中取得了很好的成绩，展示了其在大规模图像识别任务中的有效性。
以下是VGG-16的主要特点：
1. 深度：VGG-16由16个卷积层和3个全连接层组成，因此具有相对较深的网络结构。这种深度有助于网络学习到更加抽象和复杂的特征。
2. 卷积层的设计：VGG-16的卷积层全部采用3x3的卷积核和步长为1的卷积操作，同时在卷积层之后都接有ReLU激活函数。这种设计的好处在于，通过堆叠多个较小的卷积核，可以提高网络的非线性建模能力，同时减少了参数数量，从而降低了过拟合的风险。
3. 池化层：在卷积层之后，VGG-16使用最大池化层来减少特征图的空间尺寸，帮助提取更加显著的特征并减少计算量。
4. 全连接层：VGG-16在卷积层之后接有3个全连接层，最后一个全连接层输出与类别数相对应的向量，用于进行分类。
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
# 2.1 模型建立
from torchvision.models import vgg16
device = "cuda" if torch.cuda.is_available() else "cpu"
print("Using {} device".format(device))
# 加载预训练模型，并且对模型进行微调
model = vgg16(pretrained = True).to(device) # 加载预训练的vgg16模型
for param in model.parameters():
param.requires_grad = False # 冻结模型的参数，这样子在训练的时候只训练最后一层的参数
# 修改classifier模块的第6层（即：(6): Linear(in_features=4096, out_features=2, bias=True)）
# 注意查看我们下方打印出来的模型
model.classifier._modules['6'] = nn.Linear(4096,len(classeNames)) # 修改vgg16模型中最后一层全连接层，输出目标类别个数
model.to(device)
model
Using cpu device
Downloading: "https://download.pytorch.org/models/vgg16-397923af.pth" to /Users/corawoo/.cache/torch/hub/checkpoints/vgg16-397923af.pth 100%|█████████████████████████████████████████| 528M/528M [35:10<00:00, 262kB/s]
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=0, bias=True)
)
)
# 三、训练模型
# 3.1 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
# 3.2 编写测试函数
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
# 3.3 设置动态学习率
def adjust_learning_rate(optimizer, epoch, start_lr):
# 每 2 个 epoch 衰减到原来的 0.98
lr = start_lr * (0.92 ** (epoch // 2))
for param_group in optimizer.param_groups:
param_group['lr'] = lr
learn_rate = 1e-4 # 初始学习率
optimizer = torch.optim.SGD(model.parameters(), lr=learn_rate)
# 打印训练数据的标签值
for batch in train_dl:
X, y = batch
print("Labels:", y)
Labels: tensor([16, 0, 1, 10, 7, 4, 6, 8, 8, 6, 2, 16, 11, 11, 11, 5, 14, 11,
5, 6, 7, 9, 7, 10, 3, 15, 13, 0, 6, 16, 13, 16])
Labels: tensor([ 3, 4, 0, 3, 14, 11, 14, 4, 15, 6, 1, 9, 16, 13, 0, 9, 0, 5,
13, 3, 7, 3, 5, 3, 13, 5, 14, 8, 5, 12, 11, 2])
Labels: tensor([ 7, 14, 14, 6, 16, 9, 2, 14, 13, 2, 3, 11, 11, 1, 8, 11, 8, 6,
9, 3, 0, 15, 13, 5, 12, 10, 13, 5, 14, 13, 1, 11])
Labels: tensor([16, 5, 1, 16, 5, 5, 0, 9, 16, 13, 2, 14, 6, 14, 7, 7, 2, 15,
6, 11, 0, 4, 11, 3, 4, 6, 1, 12, 6, 0, 13, 8])
Labels: tensor([ 9, 14, 13, 9, 8, 8, 12, 16, 7, 3, 10, 2, 14, 6, 6, 3, 3, 6,
13, 0, 13, 11, 14, 7, 5, 7, 3, 7, 7, 10, 5, 9])
Labels: tensor([ 8, 1, 2, 10, 16, 6, 7, 8, 1, 5, 4, 13, 8, 8, 7, 2, 2, 5,
2, 5, 2, 4, 13, 5, 2, 12, 15, 16, 16, 9, 5, 6])
Labels: tensor([13, 14, 15, 1, 13, 4, 3, 13, 3, 9, 13, 10, 8, 7, 12, 7, 2, 0,
8, 6, 2, 11, 1, 15, 5, 1, 1, 9, 7, 0, 15, 3])
Labels: tensor([ 1, 1, 5, 2, 13, 12, 14, 13, 12, 9, 13, 6, 9, 9, 7, 3, 4, 6,
7, 7, 10, 9, 16, 12, 11, 16, 14, 3, 2, 4, 14, 2])
Labels: tensor([ 9, 2, 4, 9, 14, 5, 15, 14, 9, 13, 2, 15, 9, 8, 6, 4, 15, 10,
15, 16, 0, 12, 8, 13, 0, 3, 4, 13, 10, 5, 5, 13])
Labels: tensor([15, 0, 1, 3, 12, 0, 10, 13, 10, 11, 7, 3, 2, 3, 6, 1, 16, 12,
10, 14, 0, 15, 11, 4, 7, 2, 14, 12, 10, 4, 16, 7])
Labels: tensor([15, 1, 4, 15, 5, 1, 11, 0, 10, 9, 11, 14, 4, 4, 4, 4, 14, 15,
11, 1, 4, 8, 15, 13, 7, 1, 14, 3, 4, 15, 14, 16])
Labels: tensor([ 2, 8, 5, 10, 7, 13, 13, 1, 2, 5, 8, 15, 12, 14, 0, 13, 2, 2,
14, 12, 2, 7, 13, 4, 10, 5, 1, 10, 6, 9, 4, 10])
Labels: tensor([ 8, 14, 3, 10, 3, 13, 8, 15, 7, 0, 13, 0, 9, 15, 15, 1, 15, 6,
13, 1, 9, 11, 2, 4, 1, 16, 2, 1, 7, 0, 4, 7])
Labels: tensor([10, 14, 9, 10, 6, 4, 13, 10, 7, 14, 6, 0, 8, 12, 4, 9, 4, 16,
4, 16, 13, 8, 5, 10, 6, 11, 15, 1, 5, 3, 15, 6])
Labels: tensor([10, 9, 15, 13, 14, 4, 16, 9, 6, 16, 12, 9, 9, 9, 11, 1, 13, 4,
13, 3, 15, 12, 16, 13, 8, 14, 13, 8, 13, 1, 0, 10])
Labels: tensor([ 1, 13, 0, 12, 7, 3, 2, 7, 8, 7, 3, 16, 0, 12, 10, 2, 15, 9,
12, 8, 11, 13, 2, 13, 11, 11, 8, 16, 13, 5, 0, 5])
Labels: tensor([15, 15, 13, 14, 13, 13, 13, 5, 3, 12, 13, 3, 14, 1, 0, 8, 4, 12,
0, 6, 13, 9, 9, 3, 3, 4, 15, 16, 2, 12, 7, 14])
Labels: tensor([16, 6, 12, 9, 11, 3, 2, 8, 13, 10, 1, 7, 15, 4, 7, 9, 6, 6,
14, 3, 1, 6, 10, 6, 10, 15, 16, 14, 11, 7, 7, 4])
Labels: tensor([12, 4, 4, 8, 13, 6, 5, 3, 12, 7, 3, 1, 8, 8, 4, 15, 15, 4,
14, 7, 12, 15, 3, 11, 10, 12, 16, 5, 10, 9, 3, 11])
Labels: tensor([ 4, 15, 7, 15, 13, 13, 2, 6, 13, 9, 11, 14, 4, 5, 6, 2, 3, 2,
0, 6, 4, 15, 12, 10, 9, 10, 9, 13, 16, 11, 6, 16])
Labels: tensor([13, 6, 2, 11, 10, 3, 9, 10, 15, 5, 9, 9, 6, 6, 2, 13, 8, 15,
5, 16, 9, 5, 3, 4, 16, 4, 11, 11, 14, 15, 3, 7])
Labels: tensor([14, 0, 14, 5, 6, 16, 1, 16, 13, 12, 4, 6, 9, 4, 14, 15, 2, 13,
0, 13, 7, 13, 13, 6, 15, 9, 9, 2, 8, 0, 9, 13])
Labels: tensor([14, 10, 10, 2, 0, 8, 5, 15, 10, 3, 3, 13, 16, 11, 13, 6, 0, 16,
12, 13, 12, 8, 15, 0, 13, 12, 2, 5, 2, 14, 6, 1])
Labels: tensor([ 7, 5, 15, 13, 5, 1, 2, 15, 14, 6, 12, 4, 16, 13, 8, 10, 1, 8,
16, 4, 6, 8, 11, 12, 16, 8, 0, 7, 3, 4, 4, 16])
Labels: tensor([12, 9, 6, 14, 13, 16, 7, 16, 10, 16, 13, 9, 1, 7, 16, 6, 3, 9,
12, 4, 1, 0, 1, 7, 10, 11, 0, 11, 2, 10, 13, 6])
Labels: tensor([11, 15, 10, 7, 3, 2, 13, 7, 1, 13, 0, 7, 0, 10, 13, 13, 9, 13,
10, 14, 13, 8, 6, 10, 13, 1, 3, 5, 9, 2, 8, 10])
Labels: tensor([ 8, 2, 14, 0, 0, 4, 15, 8, 16, 13, 9, 3, 8, 12, 14, 3, 15, 11,
12, 4, 8, 13, 11, 8, 9, 3, 14, 13, 4, 12, 5, 8])
Labels: tensor([12, 12, 13, 5, 10, 11, 1, 4, 2, 13, 13, 14, 3, 1, 1, 6, 2, 7,
0, 12, 9, 12, 16, 7, 6, 4, 13, 3, 8, 13, 9, 7])
Labels: tensor([13, 13, 5, 13, 9, 2, 5, 6, 7, 14, 12, 6, 13, 1, 2, 13, 6, 1,
13, 15, 2, 13, 10, 5, 13, 15, 1, 5, 2, 1, 16, 13])
Labels: tensor([ 1, 13, 4, 15, 13, 14, 3, 16, 5, 8, 7, 15, 7, 8, 4, 13, 5, 4,
16, 4, 0, 10, 12, 2, 12, 7, 11, 3, 4, 13, 13, 15])
Labels: tensor([ 9, 12, 11, 3, 6, 0, 9, 5, 11, 9, 6, 4, 4, 16, 5, 7, 1, 13,
16, 3, 16, 14, 8, 16, 13, 5, 6, 8, 10, 0, 10, 2])
Labels: tensor([12, 0, 9, 16, 5, 9, 7, 3, 8, 13, 13, 6, 1, 10, 1, 10, 11, 1,
5, 9, 3, 10, 0, 14, 2, 12, 8, 6, 5, 9, 15, 2])
Labels: tensor([ 0, 10, 11, 9, 2, 16, 0, 15, 9, 1, 11, 3, 0, 2, 10, 12, 9, 10,
10, 5, 4, 13, 15, 1, 7, 16, 8, 5, 9, 11, 3, 5])
Labels: tensor([15, 3, 0, 10, 8, 8, 12, 15, 16, 8, 3, 6, 7, 14, 3, 12, 10, 0,
12, 14, 15, 13, 12, 0, 13, 13, 9, 13, 3, 11, 7, 9])
Labels: tensor([12, 9, 10, 13, 2, 1, 8, 0, 12, 7, 6, 16, 16, 8, 12, 7, 15, 7,
3, 1, 13, 13, 12, 9, 0, 4, 8, 6, 5, 9, 7, 3])
Labels: tensor([13, 10, 3, 4, 14, 14, 16, 2, 11, 11, 13, 2, 5, 9, 8, 14, 8, 0,
13, 2, 9, 16, 1, 13, 3, 9, 4, 7, 5, 7, 14, 8])
Labels: tensor([13, 13, 4, 10, 2, 16, 16, 15, 13, 5, 16, 9, 7, 7, 0, 13, 5, 16,
16, 14, 13, 12, 2, 1, 0, 16, 7, 16, 9, 11, 12, 14])
Labels: tensor([16, 11, 14, 5, 7, 5, 1, 6, 15, 15, 4, 11, 12, 1, 6, 13, 4, 6,
2, 10, 13, 2, 3, 8, 13, 12, 0, 9, 5, 10, 9, 14])
Labels: tensor([13, 1, 14, 12, 12, 8, 5, 9, 12, 4, 9, 3, 4, 13, 14, 14, 9, 8,
0, 9, 2, 11, 7, 13, 4, 13, 1, 11, 1, 0, 11, 9])
Labels: tensor([11, 0, 2, 2, 11, 16, 6, 6, 1, 10, 11, 4, 12, 1, 2, 3, 12, 0,
2, 10, 14, 1, 10, 6, 1, 1, 8, 13, 6, 0, 3, 5])
Labels: tensor([16, 7, 15, 2, 5, 14, 13, 9, 5, 7, 14, 11, 2, 5, 0, 6, 12, 0,
10, 9, 4, 13, 13, 14, 0, 12, 14, 15, 6, 14, 3, 14])
Labels: tensor([ 8, 11, 11, 9, 5, 7, 15, 14, 11, 6, 5, 7, 15, 2, 12, 1, 2, 11,
14, 2, 4, 16, 1, 11, 11, 4, 15, 10, 12, 13, 6, 1])
Labels: tensor([13, 2, 13, 3, 6, 0, 13, 1, 7, 10, 12, 16, 6, 12, 3, 11, 5, 1,
13, 10, 12, 9, 15, 13, 3, 3, 4, 9, 13, 9, 3, 8])
Labels: tensor([ 8, 15, 1, 13, 12, 10, 16, 0, 0, 5, 11, 3, 4, 13, 11, 14, 0, 13,
5, 7, 12, 10, 15, 2, 0, 13, 15, 8, 8, 15, 2, 8])
Labels: tensor([ 5, 5, 1, 0, 15, 6, 5, 1, 8, 16, 11, 6, 13, 5, 11, 8, 4, 12,
0, 3, 8, 14, 14, 5, 3, 13, 13, 13, 6, 12, 15, 15])
# 3.4 正式训练
import copy
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
# 定义学习率调度器
scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=10, gamma=0.5)
for epoch in range(epochs):
# 更新学习率（使用自定义学习率时使用）
# adjust_learning_rate(optimizer, epoch, learn_rate)
model.train()
epoch_train_acc, epoch_train_loss = train(train_dl, model, loss_fn, optimizer)
scheduler.step() # 更新学习率（调用官方动态学习率接口时使用）
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 = '/Users/corawoo/Desktop/data_vgg16' # 保存的参数文件名
torch.save(model.state_dict(), PATH)
print('Done')
Epoch: 1, Train_acc:7.4%, Train_loss:2.895, Test_acc:15.0%, Test_loss:2.774, Lr:1.00E-04 Epoch: 2, Train_acc:8.3%, Train_loss:2.873, Test_acc:15.3%, Test_loss:2.757, Lr:1.00E-04 Epoch: 3, Train_acc:11.4%, Train_loss:2.822, Test_acc:16.4%, Test_loss:2.730, Lr:1.00E-04 Epoch: 4, Train_acc:12.5%, Train_loss:2.789, Test_acc:15.6%, Test_loss:2.705, Lr:1.00E-04 Epoch: 5, Train_acc:13.5%, Train_loss:2.757, Test_acc:16.7%, Test_loss:2.685, Lr:1.00E-04 Epoch: 6, Train_acc:12.2%, Train_loss:2.738, Test_acc:17.8%, Test_loss:2.661, Lr:1.00E-04 Epoch: 7, Train_acc:14.8%, Train_loss:2.711, Test_acc:17.8%, Test_loss:2.645, Lr:1.00E-04 Epoch: 8, Train_acc:13.6%, Train_loss:2.703, Test_acc:18.1%, Test_loss:2.620, Lr:1.00E-04 Epoch: 9, Train_acc:14.6%, Train_loss:2.686, Test_acc:18.3%, Test_loss:2.620, Lr:1.00E-04 Epoch:10, Train_acc:14.0%, Train_loss:2.664, Test_acc:18.6%, Test_loss:2.599, Lr:5.00E-05 Epoch:11, Train_acc:15.2%, Train_loss:2.655, Test_acc:18.6%, Test_loss:2.587, Lr:5.00E-05 Epoch:12, Train_acc:15.2%, Train_loss:2.648, Test_acc:18.1%, Test_loss:2.585, Lr:5.00E-05 Epoch:13, Train_acc:14.3%, Train_loss:2.640, Test_acc:18.3%, Test_loss:2.582, Lr:5.00E-05 Epoch:14, Train_acc:16.0%, Train_loss:2.624, Test_acc:18.9%, Test_loss:2.560, Lr:5.00E-05 Epoch:15, Train_acc:16.2%, Train_loss:2.620, Test_acc:19.2%, Test_loss:2.555, Lr:5.00E-05 Epoch:16, Train_acc:16.5%, Train_loss:2.609, Test_acc:19.4%, Test_loss:2.550, Lr:5.00E-05 Epoch:17, Train_acc:15.7%, Train_loss:2.603, Test_acc:19.2%, Test_loss:2.523, Lr:5.00E-05 Epoch:18, Train_acc:17.2%, Train_loss:2.580, Test_acc:19.2%, Test_loss:2.528, Lr:5.00E-05 Epoch:19, Train_acc:17.2%, Train_loss:2.590, Test_acc:19.4%, Test_loss:2.529, Lr:5.00E-05 Epoch:20, Train_acc:18.3%, Train_loss:2.556, Test_acc:19.4%, Test_loss:2.510, Lr:2.50E-05 Epoch:21, Train_acc:16.4%, Train_loss:2.573, Test_acc:19.4%, Test_loss:2.520, Lr:2.50E-05 Epoch:22, Train_acc:17.6%, Train_loss:2.575, Test_acc:19.4%, Test_loss:2.521, Lr:2.50E-05 Epoch:23, Train_acc:17.6%, Train_loss:2.563, Test_acc:20.0%, Test_loss:2.513, Lr:2.50E-05 Epoch:24, Train_acc:15.8%, Train_loss:2.559, Test_acc:20.0%, Test_loss:2.490, Lr:2.50E-05 Epoch:25, Train_acc:18.7%, Train_loss:2.544, Test_acc:20.0%, Test_loss:2.515, Lr:2.50E-05 Epoch:26, Train_acc:19.7%, Train_loss:2.532, Test_acc:19.7%, Test_loss:2.502, Lr:2.50E-05 Epoch:27, Train_acc:17.7%, Train_loss:2.543, Test_acc:20.3%, Test_loss:2.497, Lr:2.50E-05 Epoch:28, Train_acc:17.9%, Train_loss:2.544, Test_acc:20.3%, Test_loss:2.503, Lr:2.50E-05 Epoch:29, Train_acc:18.1%, Train_loss:2.550, Test_acc:20.3%, Test_loss:2.487, Lr:2.50E-05 Epoch:30, Train_acc:18.7%, Train_loss:2.540, Test_acc:20.3%, Test_loss:2.490, Lr:1.25E-05 Epoch:31, Train_acc:16.8%, Train_loss:2.538, Test_acc:20.3%, Test_loss:2.491, Lr:1.25E-05 Epoch:32, Train_acc:17.4%, Train_loss:2.536, Test_acc:20.3%, Test_loss:2.492, Lr:1.25E-05 Epoch:33, Train_acc:17.4%, Train_loss:2.553, Test_acc:20.0%, Test_loss:2.484, Lr:1.25E-05 Epoch:34, Train_acc:18.7%, Train_loss:2.541, Test_acc:20.3%, Test_loss:2.458, Lr:1.25E-05 Epoch:35, Train_acc:20.1%, Train_loss:2.527, Test_acc:20.3%, Test_loss:2.493, Lr:1.25E-05 Epoch:36, Train_acc:18.7%, Train_loss:2.518, Test_acc:20.3%, Test_loss:2.494, Lr:1.25E-05 Epoch:37, Train_acc:17.4%, Train_loss:2.532, Test_acc:20.3%, Test_loss:2.477, Lr:1.25E-05 Epoch:38, Train_acc:17.4%, Train_loss:2.532, Test_acc:20.3%, Test_loss:2.475, Lr:1.25E-05 Epoch:39, Train_acc:19.0%, Train_loss:2.506, Test_acc:20.3%, Test_loss:2.481, Lr:1.25E-05 Epoch:40, Train_acc:18.3%, Train_loss:2.519, Test_acc:20.3%, Test_loss:2.465, Lr:6.25E-06 Done
# 四、结果可视化
# 4.1 loss和accuracy图
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()
# 4.2 指定图片进行预测
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}')
# 预测训练集中的某张照片
predict_one_image(image_path='/Users/corawoo/Desktop/data6/Angelina Jolie/001_fe3347c0.jpg',
model=model,
transform=train_transforms,
classes=classes)
# 4.3 模型评估
best_model.eval()
epoch_test_acc, epoch_test_loss = test(test_dl, best_model, loss_fn)
epoch_test_acc, epoch_test_loss
(0.20277777777777778, 2.4938654700915017)
# 查看是否与我们记录的最高准确率一致
epoch_test_acc
0.20277777777777778
DAMO开发者矩阵，由阿里巴巴达摩院和中国互联网协会联合发起，致力于探讨最前沿的技术趋势与应用成果，搭建高质量的交流与分享平台，推动技术创新与产业应用链接，围绕“人工智能与新型计算”构建开放共享的开发者生态。
更多推荐
41
0- 0
已为社区贡献1条内容
所有评论(0)