pytorch小土堆

无敌小短腿

已于 2023-04-09 21:21:48 修改

阅读量841

点赞数 1

分类专栏：深度学习文章标签： pytorch 学习 python

于 2023-03-12 23:40:10 首次发布

原文链接：https://www.bilibili.com/video/BV1hE411t7RN/?p=7&spm_id_from=pageDriver&vd_source=1802b9c2c0559adcc97f13025eb9398c

版权

深度学习专栏收录该内容

5 篇文章 0 订阅

订阅专栏

深度学习代码

一、数据集部分

Dataset、DataLoader、transforms、SummaryWriter

二、建模部分

torch.nn包：Linear()、Conv2d()、MaxPool2d() 、Sigmoid()、ReLU()、BatchNorm2d()、Flatten()、Sequential()

三、前向传播、反向传播部分

损失器、优化器

训练：前向、反向、训练

from torch.nn import Conv2d, MaxPool2d, Flatten, Linear, Sequential
Sequential(
            Conv2d(3, 32, 5, padding=2),#输入通道、输出通道、卷积核、padding
            MaxPool2d(2),#池化
            Conv2d(32, 32, 5, padding=2),
            MaxPool2d(2),
            Conv2d(32, 64, 5, padding=2),
            MaxPool2d(2),
            Flatten(),#展平
            Linear(1024, 64),#输入特征、输出特征
            Linear(64, 10)
        )

数据集

一、dir()与help()

dir()函数与help()函数
dir() 查看包里有什么
help() 查看怎么使用 help(Dataset)
Dataset?? 查看怎么使用
ctrl按住类查看
![在这里插入图片描述](https://img-blog.csdnimg.cn/c96ab9c59d4c4c43b789283a88ce9d61.png)

二、python文件、控制台、jupyter区别

三、Dataset与DataLoader类

Dataset

1.创建一个myData类，继承Dataset类
2.重要函数
(1)初始化函数__init__():得到所有样本以及标签
(2)获取样本对函数__getitem__():获取样本对，模型直接通过这一函数获得一对样本对{x:y}

(3)返回数据集长度__len__()

3.路径函数
(1)os.path.join()将多个路径拼接
(2)os.listdir()将路径下的所有文件名组成一个列表
```

4.Dataset对象相加，即可使得多个数据集拼接；Dataset对象中可以使用root_dir和label_dir找到同标签的所有数据集

DataLoader

test_loader = DataLoader(dataset=test_data, batch_size=64, shuffle=True, num_workers=0, drop_last=True)

#writer.add_images可以一幅图输出多张图片

import torchvision

# 准备的测试数据集
from torch.utils.data import DataLoader
from torch.utils.tensorboard import SummaryWriter

test_data = torchvision.datasets.CIFAR10("./dataset", train=False, transform=torchvision.transforms.ToTensor())

test_loader = DataLoader(dataset=test_data, batch_size=64, shuffle=True, num_workers=0, drop_last=True)

# 测试数据集中第一张图片及target
img, target = test_data[0]
print(img.shape)
print(target)

writer = SummaryWriter("dataloader")
for epoch in range(2):
    step = 0
    for data in test_loader:
        imgs, targets = data
        # print(imgs.shape)
        # print(targets)
        writer.add_images("Epoch: {}".format(epoch), imgs, step)
        step = step + 1

writer.close()

四、tensorboard

1.SummaryWriter类

能够绘制train_loss的图像，保存在logs文件夹里

（1）代码：

#coding=utf-8
'''
1.SummaryWriter类
writer=SummaryWriter("logs") #logs指的是文件夹
writer.add_scalar("y=2x", 2*i, i) #图像名，描点连直线画图

writer.add_image("train", img_array, 1, dataformats='HWC') #train是图像名；img_array必须是numpy或者tensor型；1指的是step，把step改成2再次运行后，train图像可以左右拖动
'''
from torch.utils.tensorboard import SummaryWriter
import numpy as np
from PIL import Image

writer = SummaryWriter("logs")
image_path = "../hymenoptera_data/train/bees/205835650_e6f2614bee.jpg"
image_path1 = "../hymenoptera_data/train/ants/0013035.jpg"
img_PIL = Image.open(image_path)
img_array = np.array(img_PIL)
print(type(img_array))
print(img_array.shape)#可以看出是(H,W,C)型
img_PIL1 = Image.open(image_path1)
img_array1 = np.array(img_PIL1)
#
#logs文件夹中保存训练的数据集
writer.add_image("train", img_array, 1, dataformats='HWC')
writer.add_image("train", img_array1, 2, dataformats='HWC')

#logs文件中保存画图信息
# y = 2x
for i in range(100):
    writer.add_scalar("y=2x", 2*i, i)#"y=2x"对应一副图像，再来一次不改变""里图像名的话会在原图像上画，线会叠加

writer.close()

（2）同路径的终端下输入（其中port是指定服务器，默认是6006，若冲突则更换）：

tensorboard --logdir=logs --port=6007

点击后显示：

2.call函数可以使得对象当函数用

注意

五、tranforms

1.transforms用法

2.ToTensor()使用

T=transforms.ToTensor()  #创建一个对象

input_tensor=T(input)#将numpy转化为tensor对象，将hwc转为chw

3.Normalize()使用

标准化，将像素值映射到-1到1之间

4.Resize()使用

改变图像大小，输入的是PIL或者tensor，输入与输出格式对应

5.Compose()使用

将多个操作组合在一起，注意前面对象的输出一定要对应后面对象的输入

transform = transforms.Compose([transforms.Resize(400), transforms.ToTensor()])

建模

一、torch.nn包

二、torch.nn.Module类

1.任何模型需要继承Module类，并且实现init()以及forward()

2.调试代码

三、torch.nn.Cov2d()函数

1.卷积：对于3通道图像，卷积核通道数必须等于图像通道数

最原始的图像数据类型应该是BCHW类型的数据，卷积操作不改变B

 input=torch.reshape(input,(1,1,2,2))#将input转化为BCHW类型，作为输入层

所以计算CWH变化时，先对一个样本进行计算

conv1 = torch.nn.Conv2d(in_channels=3, out_channels=6, kernel_size=3, stride=1, padding=0)

四、torch.nn.MaxPool2d()函数——下采样，即池化

最大池化的目的是保留原始数据特征情况下，减小数据量

maxpool1 = torch.nn.MaxPool2d(kernel_size=3, ceil_mode=False)

五、非线性激活函数

1.torch.nn.Sigmoid()

2.torch.nn.ReLU()

六、正则化层——可以加快网络训练速度

m=torch.nn.BatchNorm2d(100,affine=False)#正则化对象
input=torch.randn(20,100,35,45)#BCHW类型数据
output=m(input)#将input正则化

七、线性层

torch.nn.Flatten(img)#将张量展平为一维的
linear1=torch.nn.Linear(in_features,out_features)#线性层改变特征数量

八、Sequential函数

1.简单的网络结构

import torch
from torch import nn
from torch.nn import Conv2d, MaxPool2d, Flatten, Linear, Sequential
from torch.utils.tensorboard import SummaryWriter

class Tudui(nn.Module):
    def __init__(self):
        super(Tudui, self).__init__()
        self.model1 = Sequential(
            Conv2d(3, 32, 5, padding=2),
            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

tudui = Tudui()#生成一个模型对象
input = torch.ones((64, 3, 32, 32))#输入数据，必须是(B,C,H,W)类型
output = tudui(input)
print(output.shape)

#可视化构造图
writer = SummaryWriter("../logs_seq")
writer.add_graph(tudui, input)
writer.close()

模型可视化结果：

前向传播与反向传播

一、前向传播求得y_hat以及loss


#损失器对象
loss = L1Loss(reduction='sum')
loss_mse = nn.MSELoss()
loss_cross = nn.CrossEntropyLoss()#交叉熵

#计算损失
l=loss(y,y_hat)

二、反向传播

loss.backward()#求解计算图内所有的梯度

三、优化器——梯度置0，梯度更新

四、训练

import torch
import torchvision
from torch import nn
from torch.nn import Sequential, Conv2d, MaxPool2d, Flatten, Linear
from torch.optim.lr_scheduler import StepLR
from torch.utils.data import DataLoader

dataset = torchvision.datasets.CIFAR10("../data", train=False, transform=torchvision.transforms.ToTensor(),
                                       download=True)

dataloader = DataLoader(dataset, batch_size=1)

class Tudui(nn.Module):
    def __init__(self):
        super(Tudui, self).__init__()
        self.model1 = Sequential(
            Conv2d(3, 32, 5, padding=2),
            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


loss = nn.CrossEntropyLoss()
tudui = Tudui()
optim = torch.optim.SGD(tudui.parameters(), lr=0.01)
for epoch in range(20):
    running_loss = 0.0
    for data in dataloader:
        imgs, targets = data
        outputs = tudui(imgs)
        result_loss = loss(outputs, targets)
        optim.zero_grad()
        result_loss.backward()
        optim.step()
        running_loss = running_loss + result_loss
    print(running_loss)

现有模型的使用、修改、保存、加载

一、添加与修改

二、保存

vgg16 = torchvision.models.vgg16(pretrained=False)
# 保存方式1,模型结构+模型参数
torch.save(vgg16, "vgg16_method1.pth")

# 保存方式2，模型参数（官方推荐）
torch.save(vgg16.state_dict(), "vgg16_method2.pth")

三、加载

#方式1加载模型
model = torch.load('tudui_method1.pth')
print(model)

#方式2
vgg16 = torchvision.models.vgg16(pretrained=False)
vgg16.load_state_dict(torch.load("vgg16_method2.pth"))
print(vgg16)

完整的模型训练、测试

# -*- coding: utf-8 -*-

import torchvision
from torch.utils.tensorboard import SummaryWriter

from model import *
# 准备数据集
from torch import nn
from torch.utils.data import DataLoader

train_data = torchvision.datasets.CIFAR10(root="../data", train=True, transform=torchvision.transforms.ToTensor(),
                                          download=True)
test_data = torchvision.datasets.CIFAR10(root="../data", train=False, transform=torchvision.transforms.ToTensor(),
                                         download=True)

# length 长度
train_data_size = len(train_data)
test_data_size = len(test_data)
# 如果train_data_size=10, 训练数据集的长度为：10
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)

# 创建网络模型
tudui = Tudui()

# 损失函数
loss_fn = nn.CrossEntropyLoss()

# 优化器
# learning_rate = 0.01
# 1e-2=1 x (10)^(-2) = 1 /100 = 0.01
learning_rate = 1e-2
optimizer = torch.optim.SGD(tudui.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 = 10

# 添加tensorboard
writer = SummaryWriter("../logs_train")

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

    # 训练步骤开始
    tudui.train()
    for data in train_dataloader:
        imgs, targets = data
        outputs = tudui(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:
            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
    tudui.eval()
    total_test_loss = 0
    total_accuracy = 0
    with torch.no_grad():
        for data in test_dataloader:
            imgs, targets = data
            outputs = tudui(imgs)
            loss = loss_fn(outputs, targets)
            total_test_loss = total_test_loss + loss.item()
            accuracy = (outputs.argmax(1) == targets).sum()
            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

    torch.save(tudui, "tudui_{}.pth".format(i))
    print("模型已保存")

writer.close()

使用GPU训练

2.方式1


import torch
import torchvision
from torch.utils.tensorboard import SummaryWriter

# from model import *
# 准备数据集
from torch import nn
from torch.utils.data import DataLoader

train_data = torchvision.datasets.CIFAR10(root="../data", train=True, transform=torchvision.transforms.ToTensor(),
                                          download=True)
test_data = torchvision.datasets.CIFAR10(root="../data", train=False, transform=torchvision.transforms.ToTensor(),
                                         download=True)

# length 长度
train_data_size = len(train_data)
test_data_size = len(test_data)
# 如果train_data_size=10, 训练数据集的长度为：10
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)

# 创建网络模型
class Tudui(nn.Module):
    def __init__(self):
        super(Tudui, self).__init__()
        self.model = nn.Sequential(
            nn.Conv2d(3, 32, 5, 1, 2),
            nn.MaxPool2d(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(64*4*4, 64),
            nn.Linear(64, 10)
        )

    def forward(self, x):
        x = self.model(x)
        return x
tudui = Tudui()
if torch.cuda.is_available():     #####模型cuda
    tudui = tudui.cuda()

# 损失函数
loss_fn = nn.CrossEntropyLoss()
if torch.cuda.is_available():     #####损失函数cuda
    loss_fn = loss_fn.cuda()
# 优化器
# learning_rate = 0.01
# 1e-2=1 x (10)^(-2) = 1 /100 = 0.01
learning_rate = 1e-2
optimizer = torch.optim.SGD(tudui.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 = 10

# 添加tensorboard
writer = SummaryWriter("../logs_train")

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

    # 训练步骤开始
    tudui.train()
    for data in train_dataloader:
        imgs, targets = data
        if torch.cuda.is_available():     #####数据集cuda
            imgs = imgs.cuda()
            targets = targets.cuda()
        outputs = tudui(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:
            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
    tudui.eval()
    total_test_loss = 0
    total_accuracy = 0
    with torch.no_grad():
        for data in test_dataloader:
            imgs, targets = data
            if torch.cuda.is_available():      #####数据集cuda
                imgs = imgs.cuda()
                targets = targets.cuda()
            outputs = tudui(imgs)
            loss = loss_fn(outputs, targets)
            total_test_loss = total_test_loss + loss.item()
            accuracy = (outputs.argmax(1) == targets).sum()
            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

    torch.save(tudui, "tudui_{}.pth".format(i))
    print("模型已保存")

writer.close()

3.方式2

import torch
import torchvision
from torch.utils.tensorboard import SummaryWriter

# from model import *
# 准备数据集
from torch import nn
from torch.utils.data import DataLoader

# 定义训练的设备
device = torch.device("cuda")

train_data = torchvision.datasets.CIFAR10(root="../data", train=True, transform=torchvision.transforms.ToTensor(),
                                          download=True)
test_data = torchvision.datasets.CIFAR10(root="../data", train=False, transform=torchvision.transforms.ToTensor(),
                                         download=True)

# length 长度
train_data_size = len(train_data)
test_data_size = len(test_data)
# 如果train_data_size=10, 训练数据集的长度为：10
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)

# 创建网络模型
class Tudui(nn.Module):
    def __init__(self):
        super(Tudui, self).__init__()
        self.model = nn.Sequential(
            nn.Conv2d(3, 32, 5, 1, 2),
            nn.MaxPool2d(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(64*4*4, 64),
            nn.Linear(64, 10)
        )

    def forward(self, x):
        x = self.model(x)
        return x
tudui = Tudui()
tudui = tudui.to(device)

# 损失函数
loss_fn = nn.CrossEntropyLoss()
loss_fn = loss_fn.to(device)
# 优化器
# learning_rate = 0.01
# 1e-2=1 x (10)^(-2) = 1 /100 = 0.01
learning_rate = 1e-2
optimizer = torch.optim.SGD(tudui.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 = 10

# 添加tensorboard
writer = SummaryWriter("../logs_train")

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

    # 训练步骤开始
    tudui.train()
    for data in train_dataloader:
        imgs, targets = data
        imgs = imgs.to(device)
        targets = targets.to(device)
        outputs = tudui(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:
            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
    tudui.eval()
    total_test_loss = 0
    total_accuracy = 0
    with torch.no_grad():
        for data in test_dataloader:
            imgs, targets = data
            imgs = imgs.to(device)
            targets = targets.to(device)
            outputs = tudui(imgs)
            loss = loss_fn(outputs, targets)
            total_test_loss = total_test_loss + loss.item()
            accuracy = (outputs.argmax(1) == targets).sum()
            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

    torch.save(tudui, "tudui_{}.pth".format(i))
    print("模型已保存")

writer.close()