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
我的环境:
1.语言:python3.7
2.编译器:pycharm
一、前期准备
1、设置GPU
import torch
import torch.nn as nn
import matplotlib.pyplot as plt
import torchvision
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
device
2、导入数据
import os,PIL,random,pathlib
data_dir = "E:\TF环境\天气识别\weather_photos"
data_dir = pathlib.Path(data_dir)
data_paths = list(data_dir.glob('*'))
classNames = [str(path).split('\\')[1] for path in data_paths]
classNames
-
使用
pathlib.Path()
函数将字符串类型的文件夹路径转换为pathlib.Path
对象,以便后续操作。 -
使用
glob('*')
方法获取了data_dir
目录下所有文件和文件夹的路径。'*'
通配符代表匹配任意文件或文件夹名。 -
使用列表推导式遍历
data_paths
列表中的每个路径,通过将路径转换为字符串并以''为分隔符,然后取第二个元素作为类别名称,并将其添加到classNames
列表中。
['cloudy', 'rain', 'shine', 'sunrise']
total_datadir = "E:\TF环境\天气识别\weather_photos"
train_transforms = transforms.Compose([
transforms.Resize([224, 224]), # resize输入图片成统一尺寸
transforms.ToTensor(), # 将PIL Image或numpy.ndarray转换成tensor
transforms.Normalize(
mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225]) # 数据标准化处理,转换为标准正太分布
])
total_data = datasets.ImageFolder(data_dir, transform=train_transforms)
total_data
Dataset ImageFolder
Number of datapoints: 1125
Root location: E:\TF环境\天气识别\weather_photos
StandardTransform
Transform: Compose(
Resize(size=[224, 224], interpolation=bilinear, max_size=None, antialias=None)
ToTensor()
Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
3、划分训练集
train_size = int(0.8 * len(total_data))
test_size = len(total_data) - train_size#训练集测试集8 2 分
train_dataset, test_dataset = torch.utils.data.random_split(total_data, [train_size, test_size])
train_dataset, test_dataset
- 使用 torch.utils.data.random_split() 函数对 total_data 进行随机划分,在此函数中,将会根据指定的两个参数(即 [train_size, test_size])将数据集划分成两个子集,其中第一个子集为训练集(train_dataset),包含 train_size 个样本;第二个子集为测试集(test_dataset),包含 test_size 个样本。
train_size,test_size
(900, 225)
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,#这里可以不用
num_workers=1)
batch_size = 32
表示每个批次加载的样本数量为32个,即每次训练或测试的时候都会同时处理32个样本。
train_dl = torch.utils.data.DataLoader(train_dataset, batch_size=batch_size, shuffle=True, num_workers=1)
创建了一个训练集的数据加载器。train_dataset
是训练集的数据集对象,batch_size
指定每个批次加载的样本数量,shuffle=True
表示在每个epoch(整个训练集迭代一次)之前将训练集打乱顺序,num_workers=1
表示使用一个线程来加载数据。
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、构建CNN网络
对于一般的CNN网络来说,都是由特征提取网络和分类网络构成,其中特征提取网络用于提取图片的特征,分类网络用于将图片进行分类。
1. torch.nn.Conv2d()详
函数原型:
torch.nn.Conv2d(in_channels, out_channels, kernel_size, stride=1, padding=0, dilation=1, groups=1, bias=True, padding_mode='zeros', device=None, dtype=None)
关键参数说明:
- in_channels ( int ) – 输入图像中的通道数
- out_channels ( int ) – 卷积产生的通道数
- kernel_size ( int or tuple ) – 卷积核的大小
- stride ( int or tuple , optional ) -- 卷积的步幅。默认值:1
- padding ( int , tuple或str , optional ) – 添加到输入的所有四个边的填充。默认值:0
- padding_mode (字符串,可选) – 'zeros', 'reflect', 'replicate'或'circular'. 默认:'zeros'
2. torch.nn.Linear()详解
函数原型:
>torch.nn.Linear(in_features, out_features, bias=True, device=None, dtype=None)
关键参数说明:
- in_features:每个输入样本的大小
- out_features:每个输出样本的大小
3. torch.nn.MaxPool2d()详解
函数原型:
>torch.nn.MaxPool2d(kernel_size, stride=None, padding=0, dilation=1, return_indices=False, ceil_mode=False)
关键参数说明:
- kernel_size:最大的窗口大小
- stride:窗口的步幅,默认值为`kernel_size`
- padding:填充值,默认为0
- dilation:控制窗口中元素步幅的参数
4. 关于卷积层、池化层的计算:
下面的网络数据shape变化过程为:
`3, 32, 32`(输入数据)
-> `64, 30, 30`(经过卷积层1)-> `64, 15, 15`(经过池化层1)
-> `64, 13, 13`(经过卷积层2)-> `64, 6, 6`(经过池化层2)
-> `128, 4, 4`(经过卷积层3) -> `128, 2, 2`(经过池化层3)
-> `512` -> `256` -> `num_classes(10)
import torch.nn.functional as F
class Network_bn(nn.Module):
def __init__(self):
super(Network_bn, self).__init__()
"""
nn.Conv2d()函数:
第一个参数(in_channels)是输入的channel数量
第二个参数(out_channels)是输出的channel数量
第三个参数(kernel_size)是卷积核大小
第四个参数(stride)是步长,默认为1
第五个参数(padding)是填充大小,默认为0
"""
self.conv1 = nn.Conv2d(in_channels=3, out_channels=12, kernel_size=5, stride=1, padding=0)#卷积层
self.bn1 = nn.BatchNorm2d(12)#批量归一化层
self.conv2 = nn.Conv2d(in_channels=12, out_channels=12, kernel_size=5, stride=1, padding=0)
self.bn2 = nn.BatchNorm2d(12)
self.pool = nn.MaxPool2d(2,2)
self.conv4 = nn.Conv2d(in_channels=12, out_channels=24, kernel_size=5, stride=1, padding=0)
self.bn4 = nn.BatchNorm2d(24)
self.conv5 = nn.Conv2d(in_channels=24, out_channels=24, kernel_size=5, stride=1, padding=0)
self.bn5 = nn.BatchNorm2d(24)
self.fc1 = nn.Linear(24*50*50, len(classeNames))
def forward(self, x):
x = F.relu(self.bn1(self.conv1(x)))
x = F.relu(self.bn2(self.conv2(x)))
x = self.pool(x)
x = F.relu(self.bn4(self.conv4(x)))
x = F.relu(self.bn5(self.conv5(x)))
x = self.pool(x)
x = x.view(-1, 24*50*50)
x = self.fc1(x)
return x
device = "cuda" if torch.cuda.is_available() else "cpu"
print("Using {} device".format(device))
model = Network_bn().to(device)
model
Using cuda device
Network_bn(
(conv1): Conv2d(3, 12, kernel_size=(5, 5), stride=(1, 1))
(bn1): BatchNorm2d(12, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(12, 12, kernel_size=(5, 5), stride=(1, 1))
(bn2): BatchNorm2d(12, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(pool): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(conv4): Conv2d(12, 24, kernel_size=(5, 5), stride=(1, 1))
(bn4): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv5): Conv2d(24, 24, kernel_size=(5, 5), stride=(1, 1))
(bn5): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(fc1): Linear(in_features=60000, out_features=4, bias=True)
)
三、模型训练
1、设置超参数
loss_fn = nn.CrossEntropyLoss() # 创建损失函数,交叉熵损失函数
learn_rate = 1e-4 # 学习率
opt = torch.optim.SGD(model.parameters(),lr=learn_rate)
2、编写训练函数
1. optimizer.zero_grad()
函数会遍历模型的所有参数,通过内置方法截断反向传播的梯度流,再将每个参数的梯度值设为0,即上一次的梯度记录被清空。
2. loss.backward()
PyTorch的反向传播(即`tensor.backward()`)是通过autograd包来实现的,autograd包会根据tensor进行过的数学运算来自动计算其对应的梯度。
具体来说,torch.tensor是autograd包的基础类,如果你设置tensor的requires_grads为True,就会开始跟踪这个tensor上面的所有运算,如果你做完运算后使用`tensor.backward()`,所有的梯度就会自动运算,tensor的梯度将会累加到它的.grad属性里面去。
更具体地说,损失函数loss是由模型的所有权重w经过一系列运算得到的,若某个w的requires_grads为True,则w的所有上层参数(后面层的权重w)的.grad_fn属性中就保存了对应的运算,然后在使用`loss.backward()`后,会一层层的反向传播计算每个w的梯度值,并保存到该w的.grad属性中。
如果没有进行`tensor.backward()`的话,梯度值将会是None,因此`loss.backward()`要写在`optimizer.step()`之前。
3. optimizer.step()
step()函数的作用是执行一次优化步骤,通过梯度下降法来更新参数的值。因为梯度下降是基于梯度的,所以在执行`optimizer.step()`函数前应先执行`loss.backward()`函数来计算梯度。
注意:optimizer只负责通过梯度下降进行优化,而不负责产生梯度,梯度是`tensor.backward()`方法产生的。
# 训练循环
def train(dataloader, model, loss_fn, optimizer):
size = len(dataloader.dataset) # 训练集的大小,一共60000张图片
num_batches = len(dataloader) # 批次数目,1875(60000/32)
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. 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,一共10000张图片
num_batches = len(dataloader) # 批次数目,313(10000/32=312.5,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
4、正式训练
1. model.train()
model.train()
的作用是启用 Batch Normalization 和 Dropout。
如果模型中有BN
层(Batch Normalization)和Dropout
,需要在训练时添加model.train()
。model.train()
是保证BN层能够用到每一批数据的均值和方差。对于Dropout
,model.train()
是随机取一部分网络连接来训练更新参数。
2. model.eval()
model.eval()
的作用是不启用 Batch Normalization 和 Dropout。
如果模型中有BN层(Batch Normalization)和Dropout,在测试时添加model.eval()
。model.eval()
是保证BN层能够用全部训练数据的均值和方差,即测试过程中要保证BN层的均值和方差不变。对于Dropout
,model.eval()
是利用到了所有网络连接,即不进行随机舍弃神经元。
训练完train样本后,生成的模型model要用来测试样本。在model(test)
之前,需要加上model.eval()
,否则的话,有输入数据,即使不训练,它也会改变权值。这是model中含有BN层和Dropout所带来的的性质。
epochs = 20
train_loss = []
train_acc = []
test_loss = []
test_acc = []
for epoch in range(epochs):
model.train()
epoch_train_acc, epoch_train_loss = train(train_dl, model, loss_fn, opt)
model.eval()
epoch_test_acc, epoch_test_loss = test(test_dl, model, loss_fn)
train_acc.append(epoch_train_acc)
train_loss.append(epoch_train_loss)
test_acc.append(epoch_test_acc)
test_loss.append(epoch_test_loss)
template = ('Epoch:{:2d}, Train_acc:{:.1f}%, Train_loss:{:.3f}, Test_acc:{:.1f}%,Test_loss:{:.3f}')
print(template.format(epoch+1, epoch_train_acc*100, epoch_train_loss, epoch_test_acc*100, epoch_test_loss))
print('Done')
Epoch: 1, Train_acc:81.1%, Train_loss:0.587, Test_acc:83.6%,Test_loss:0.453
Epoch: 2, Train_acc:85.9%, Train_loss:0.499, Test_acc:86.7%,Test_loss:0.449
Epoch: 3, Train_acc:88.1%, Train_loss:0.449, Test_acc:82.7%,Test_loss:0.784
Epoch: 4, Train_acc:89.2%, Train_loss:0.405, Test_acc:85.3%,Test_loss:0.427
Epoch: 5, Train_acc:89.6%, Train_loss:0.389, Test_acc:88.9%,Test_loss:0.385
Epoch: 6, Train_acc:90.4%, Train_loss:0.338, Test_acc:88.4%,Test_loss:0.327
Epoch: 7, Train_acc:90.4%, Train_loss:0.346, Test_acc:88.9%,Test_loss:0.338
Epoch: 8, Train_acc:91.8%, Train_loss:0.302, Test_acc:86.7%,Test_loss:0.302
Epoch: 9, Train_acc:92.4%, Train_loss:0.275, Test_acc:89.3%,Test_loss:0.447
Epoch:10, Train_acc:92.4%, Train_loss:0.279, Test_acc:89.3%,Test_loss:0.354
Epoch:11, Train_acc:92.6%, Train_loss:0.252, Test_acc:87.6%,Test_loss:0.341
Epoch:12, Train_acc:92.7%, Train_loss:0.288, Test_acc:86.2%,Test_loss:0.297
Epoch:13, Train_acc:93.6%, Train_loss:0.245, Test_acc:88.9%,Test_loss:0.302
Epoch:14, Train_acc:94.0%, Train_loss:0.238, Test_acc:87.1%,Test_loss:0.279
Epoch:15, Train_acc:95.6%, Train_loss:0.225, Test_acc:88.4%,Test_loss:0.284
Epoch:16, Train_acc:93.1%, Train_loss:0.217, Test_acc:88.4%,Test_loss:0.273
Epoch:17, Train_acc:95.0%, Train_loss:0.216, Test_acc:89.3%,Test_loss:0.246
Epoch:18, Train_acc:95.3%, Train_loss:0.193, Test_acc:87.6%,Test_loss:0.288
Epoch:19, Train_acc:95.2%, Train_loss:0.186, Test_acc:90.2%,Test_loss:0.258
Epoch:20, Train_acc:95.0%, Train_loss:0.187, Test_acc:87.1%,Test_loss:0.262
Done
四、结果可视化
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()