Pytorch快速入门指南

常用的包

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

Pytorch

pytorch安装

准备环境

• 安装Ancona工具
• 安装python语言
• 安装pycharm工具

以上工作安装完成后，开始真正的pytorch安装之旅，别担心，很容易

1.打开Ancona Prompt创建一个pytorch新环境

conda create -n pytorch python=版本号比如3.11

后面步骤都是y同意安装

2.激活环境

同样在Ancona Prompt中继续输入如下指令

conda activate pytorch

3.去pytorch官网找到下载pytorch指令，根据个人配置进行选择

• window下一般选择Conda
• Linux下一般选择Pip

这里要区分自己电脑是否含有独立显卡，没有的选择cpu模式就行。

如果有独立显卡，那么去NVIDIA官网查看自己适合什么版本型号进行选择即可。

如果有独立显卡，在Ancona Prompt中输入如下指令，返回True即可确认安装成功。

torch.cuda.is_available()

如果没有cpu我们通过pycharm来进行判断，首先创建一个pytorch工程，如下所示：

import torch
print(torch.cuda.is_available())
print(torch.backends.cudnn.is_available())
print(torch.cuda_version)
print(torch.backends.cudnn.version())

print(torch.__version__)

是不是发现输出false， false， None, None，是不是以为错了。不，那是因为我们安装的是CPU版本的，压根就没得cuda，cudnn这个东西。我们只要检测python版本的torch（PyTorch）在就行。

ok！恭喜你成功完成安装pytroch！接下来开启你的学习之路吧！

引言：python中的两大法宝函数

• 这里1、2、3、4是分隔区

# 查看torch里面有什么
for i in range(len(dir(torch))):
    print(f"{dir(torch)[i]}")

pytorch加载数据初认识

import import torch
from torch.utils.data import Dataset

• 看看Dataset里面有什么：

Dataset代码实战

from torch.utils.data import Dataset
from PIL import Image
import os


class MyData(Dataset):
    def __init__(self, root_dir, label_dir):
        self.root_dir = root_dir
        self.label_dir = label_dir
        self.path = os.path.join(self.root_dir, self.label_dir)  # 根路径和最后的路径进行拼接
        self.img_path = os.listdir(self.path)  # 路径地址img_path[0] 就是第一张地址

    def __getitem__(self, idx):
        """
        读取每个照片
        :param idx:
        :return:
        """
        img_name = self.img_path[idx]
        img_item_path = os.path.join(self.root_dir, self.label_dir, img_name)
        img = Image.open(img_item_path)
        label = self.label_dir
        return img, label

    def __len__(self):
        """
        查看图片个数，即数据集个数
        :return:
        """
        return len(self.img_path)


# img_path = "E:\\Project\\Code_Python\\Learn_pytorch\\learn_pytorch\\dataset\\training_set\\cats\\cat.1.jpg"
# img = Image.open(img_path)
# print(img)
# img.show()

root_dir = "dataset/training_set"
cats_label_dir = "cats"
dogs_label_dir = "dogs"
cats_dataset = MyData(root_dir, cats_label_dir)
dogs_dataset = MyData(root_dir, dogs_label_dir)

img1, label1 = cats_dataset[1]
img2, label2 = dogs_dataset[1]

# img1.show()
# img2.show()

train_dataset = cats_dataset + dogs_dataset # 合并数据集
print(len(train_dataset))
print(len(cats_dataset))
print(len(dogs_dataset))

TensorBoard的使用（一）

from torch.utils.tensorboard import SummaryWriter

writer = SummaryWriter("logs")  # 事件文件存储地址
# writer.add_image()
# y = x
for i in range(100):
    writer.add_scalar("y=2x", 2*i, i)  # 标量的意思 参数2*i 是x轴  i是y轴

writer.close()

• 安装tensorboard

pip install tensorboard

• 运行tensorboard

tensorboard --logdir="logs" --port=6007（这里是指定端口号，也可以不写--port，默认6006）

• 利用Opencv读取图片，获得numpy型图片数据

• import numpy as np
  from torch.utils.tensorboard import SummaryWriter
  import cv2
  
  writer = SummaryWriter("logs")  # 事件文件存储地址
  img_array = cv2.imread("./dataset/training_set/cats/cat.2.jpg")
  # print(img_array.shape)
  writer.add_image("test",img_array,2,dataformats='HWC')
  # y = x
  for i in range(100):
      writer.add_scalar("y=2x", 2 * i, i)  # 标量的意思
  
  writer.close()
  

Transforms使用

![Snipaste_2023-11-01_10-52-09](./pytorch截图/Snipaste_2023-11-01_10-52-09.png)from torchvision import transforms
from PIL import Image

# python当中的用法
# tensor数据类型
# 通过transforms.ToTensor去解决两个问题
# 1.transforms如何使用（pyhton）
# 2.为什么需要Tensor数据类型：因为里面包装了神经网络模型训练的数据类型

# 绝对路径 E:\Project\Code_Python\Learn_pytorch\learn_pytorch\dataset\training_set\cats\cat.6.jpg
# 相对路径 dataset/training_set/cats/cat.6.jpg
img_path = "dataset/training_set/cats/cat.6.jpg"
img = Image.open(img_path)

# 1.transforms如何使用（pyhton）
tensor_trans = transforms.ToTensor()
tensor_img = tensor_trans(img)

print(tensor_img.shape)

常见的Transforms

from PIL import Image
from torchvision import transforms
from torch.utils.tensorboard import SummaryWriter

writer = SummaryWriter("logs")
img = Image.open("dataset/training_set/cats/cat.11.jpg")
print(img)

# ToTensor的使用
trans_totensor = transforms.ToTensor()
img_tensor = trans_totensor(img)
writer.add_image("ToTensor", img_tensor)

# Normalize
print(img_tensor[0][0][0])
trans_norm = transforms.Normalize([1, 1, 1], [1, 1, 1])
img_norm = trans_norm(img_tensor)
print(img_norm[0][0][0])
writer.add_image("Normalize", img_norm, 0)

# Resize
print(img.size)
trans_resize = transforms.Resize((512, 512))
# img PIL -> resize -> img_resize PIL
img_resize = trans_resize(img)
# img_resize PIL -> totensor -> img_resize tensor
img_resize = trans_totensor(img_resize)
# print(img_resize)
writer.add_image("Resize", img_resize, 1)

# Compose - resize - 2
trans_resize_2 = transforms.Resize(144)
# PIL -> PIL -> tensor数据类型
trans_compose = transforms.Compose([trans_resize_2, trans_totensor])
img_resize_2 = trans_compose(img)
writer.add_image("Resize_Compose", img_resize_2, 2)

writer.close()

torchvision中的数据集使用

import torchvision
from torch.utils.tensorboard import SummaryWriter

dataset_transforms = torchvision.transforms.Compose([
    torchvision.transforms.ToTensor()
])
# 下载数据集
train_set = torchvision.datasets.CIFAR10(root="./dataset", train=True, transform=dataset_transforms, download=True)
test_set = torchvision.datasets.CIFAR10(root="./dataset", train=False, transform=dataset_transforms, download=True)
print(test_set[0])
print(test_set.classes)

img, target = test_set[0]
print(img)
print(target)
print(test_set.classes[target])

# img.show()
writer = SummaryWriter("p10")
for i in range(10):
    img, target = test_set[i]
    writer.add_image("test_set", img, i)
writer.close()

DataLoad的使用

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

# 准备的测试数据集
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=False)

# 这里数据集是之前官网下载下来的
# 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")
step = 0
for data in test_loader:
    imgs, targets = data
    # print(imgs.shape)
    # print(targets)
    writer.add_images("test_data", imgs, step)
    step = step + 1
writer.close()

• 最后一次数据不满足64张 于是将参数设置drop_last=True

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

# 准备的测试数据集
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)

# 这里数据集是之前官网下载下来的
# 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_drop_last")
step = 0
for data in test_loader:
    imgs, targets = data
    # print(imgs.shape)
    # print(targets)
    writer.add_images("test_data", imgs, step)
    step = step + 1
writer.close()

• shuffle 使用
• True 两边图片选取不一样
• False两边图片选取一样

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

# 准备的测试数据集
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)

# 这里数据集是之前官网下载下来的
# 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("Eopch: {}".format(epoch), imgs, step)
        step = step + 1
writer.close()

神经网络的基本骨架

import torch
from torch import nn


class ConvModel(nn.Module):
    def __init__(self, *args, **kwargs) -> None:
        super().__init__(*args, **kwargs)

    def forward(self, input):
        output = input + 1
        return output


convmodel = ConvModel()
x = torch.tensor(1.0)
output = convmodel(x)
print(output)

卷积操作

import torch
import torch.nn.functional as F

# 卷积输入
input = torch.tensor([[1, 2, 0, 3, 1],
                      [0, 1, 2, 3, 1],
                      [1, 2, 1, 0, 0],
                      [5, 2, 3, 1, 1],
                      [2, 1, 0, 1, 1]])
# 卷积核
kernel = torch.tensor([[1, 2, 1],
                       [0, 1, 0],
                       [2, 1, 0]])

# 进行尺寸转换
input = torch.reshape(input, (1, 1, 5, 5))
kernel = torch.reshape(kernel, (1, 1, 3, 3))

print(input.shape)
print(kernel.shape)

output = F.conv2d(input, kernel, stride=1)
print(output)
output2 = F.conv2d(input, kernel, stride=2)
print(output2)

• padding

# padding  默认填充值是0
output3 = F.conv2d(input, kernel, stride=1, padding=1)
print(output3)

结果：

tensor([[[[ 1, 3, 4, 10, 8],
[ 5, 10, 12, 12, 6],
[ 7, 18, 16, 16, 8],
[11, 13, 9, 3, 4],
[14, 13, 9, 7, 4]]]])

神经网络-卷积层

import torch
import torchvision
from torch.utils.data import DataLoader
from torch import nn
from torch.nn import Conv2d
from torch.utils.tensorboard import SummaryWriter

dataset = torchvision.datasets.CIFAR10("./dataset", train=False, transform=torchvision.transforms.ToTensor(),
                                       download=True)
dataloader = DataLoader(dataset, batch_size=64)


class NN_Conv2d(nn.Module):
    def __init__(self):
        super().__init__()
        self.conv1 = Conv2d(in_channels=3, out_channels=6, kernel_size=3, stride=1, padding=0)

    def forward(self, x):
        x = self.conv1(x)
        return x


nn_conv2d = NN_Conv2d()
# print(nn_conv2d)

writer = SummaryWriter("./logs")
step = 0
for data in dataloader:
    imgs, targets = data
    output = nn_conv2d(imgs)
    print(f"imgs: {imgs.shape}")
    print(f"output: {output.shape}")
    # 输入的大小 torch.Size([64,3,32,32])
    writer.add_images("input", imgs, step)
    # 卷积后输出的大小 torch.Size([64,,6,30,30) --> [xxx,3,30,30]
    output = torch.reshape(output, (-1, 3, 30, 30))
    writer.add_images("output", output, step)
    step += 1

# import numpy as np
import torch
import torchvision
from torch import nn
from torch.nn import Conv2d
from torch.utils.tensorboard import SummaryWriter
import cv2
from torchvision import transforms


# 创建卷积模型
class NN_Conv2d(nn.Module):
    def __init__(self):
        super().__init__()
        self.conv1 = Conv2d(in_channels=3, out_channels=6, kernel_size=3, stride=1, padding=1)

    def forward(self, x):
        x = self.conv1(x)
        return x


nn_conv2d = NN_Conv2d()
writer = SummaryWriter('logs_test')
input_img = cv2.imread("dataset/ice.jpg")

# 转化为tensor类型
trans_tensor = transforms.ToTensor()
input_img = trans_tensor(input_img)
# 设置input输入大小
input_img = torch.reshape(input_img, (-1, 3, 1312, 2100))
print(input_img.shape)
writer.add_images("input_img", input_img, 1)
# 进行卷积输出
output = nn_conv2d(input_img)
output = torch.reshape(output, (-1, 3, 1312, 2100))
print(output.shape)

writer.add_images('output_test', output, 1)
writer.close()

神经网络-最大池化

import torch
from torch import nn
from torch.nn import MaxPool2d

input_img = torch.tensor([[1, 2, 0, 3, 1],
                          [0, 1, 2, 3, 1],
                          [1, 2, 1, 0, 0],
                          [5, 2, 3, 1, 1],
                          [2, 1, 0, 1, 1]], dtype=torch.float32)
input_img = torch.reshape(input_img, (-1, 1, 5, 5))
print(input_img.shape)


# 简单的搭建卷积神经网络
class Nn_Conv_Maxpool(nn.Module):
    def __init__(self):
        super().__init__()
        self.maxpool1 = MaxPool2d(kernel_size=3, ceil_mode=True)

    def forward(self, input_img):
        output = self.maxpool1(input_img)
        return output


nn_conv_maxpool = Nn_Conv_Maxpool()
output = nn_conv_maxpool(input_img)

print(output)

import torch
import torchvision
from torch import nn
from torch.nn import MaxPool2d
from torch.utils.data import DataLoader
from torch.utils.tensorboard import SummaryWriter

dataset = torchvision.datasets.CIFAR10('./dataset', train=False, download=True,
                                       transform=torchvision.transforms.ToTensor())
dataloader = DataLoader(dataset, batch_size=64)


# 简单的搭建卷积神经网络
class Nn_Conv_Maxpool(nn.Module):
    def __init__(self):
        super().__init__()
        self.maxpool1 = MaxPool2d(kernel_size=3, ceil_mode=True)

    def forward(self, input_img):
        output = self.maxpool1(input_img)
        return output


nn_conv_maxpool = Nn_Conv_Maxpool()

writer = SummaryWriter('logs_maxpool')
step = 0
for data in dataloader:
    imgs, targets = data
    writer.add_images('input', imgs, step)
    output = nn_conv_maxpool(imgs)
    writer.add_images('output', output, step)
    step += 1

writer.close()

神经网络-非线性激活

• ReLU

import torch
from torch import nn
from torch.nn import ReLU

input = torch.tensor([[1, -0.5],
                      [-1, 3]])
input = torch.reshape(input, (-1, 1, 2, 2))
print(input.shape)


class Nn_Network_Relu(nn.Module):
    def __init__(self):
        super().__init__()
        self.relu1 = ReLU()

    def forward(self, input):
        output = self.relu1(input)
        return output


nn_relu = Nn_Network_Relu()
output = nn_relu(input)
print(outputz)

• 使用图片进行演示

import torch
import torchvision
from torch import nn
from torch.nn import ReLU
from torch.nn import Sigmoid
from torch.utils.data import DataLoader
from torch.utils.tensorboard import SummaryWriter

input = torch.tensor([[1, -0.5],
                      [-1, 3]])
input = torch.reshape(input, (-1, 1, 2, 2))
print(input.shape)

dataset = torchvision.datasets.CIFAR10('./dataset', train=False, download=True,
                                       transform=torchvision.transforms.ToTensor())
dataloader = DataLoader(dataset, batch_size=64)


class Nn_Network_Relu(nn.Module):
    def __init__(self):
        super().__init__()
        self.relu1 = ReLU()
        self.sigmoid1 = Sigmoid()

    def forward(self, input):
        output = self.sigmoid1(input)
        return output


nn_relu = Nn_Network_Relu()
nn_sigmoid = Nn_Network_Relu()

writer = SummaryWriter('logs_sigmoid')
step = 0
for data in dataloader:
    imgs, targets = data
    writer.add_images("input_imgs", imgs, step)
    output = nn_sigmoid(imgs)
    writer.add_images("output", output, step)
    step += 1

writer.close()

神经网络-线性层及其他层

• 线性层（linear layer）通常也被称为全连接层（fully connected layer）。在深度学习模型中，线性层和全连接层指的是同一种类型的神经网络层，它将输入数据与权重相乘并加上偏置，然后通过一个非线性激活函数输出结果。可以实现特征提取、降维等功能。
• 以VGG16网络模型为例，全连接层共有3层，分别是4096-4096-1000，这里的1000为ImageNet中数据集类别的数量。

import torch
import torchvision
from torch.utils.data import DataLoader
from torch import nn
from torch.nn import Linear

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

dataloader = DataLoader(dataset, batch_size=64)


class Nn_LinearModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = Linear(196608, 10)

    def forward(self, input):
        output = self.linear1(input)
        return output


nn_linearmodel = Nn_LinearModel()
for data in dataloader:
    imgs, targets = data
    print(imgs.shape)
    output = torch.flatten(imgs)
    print(output.shape)
    output = nn_linearmodel(output)
    print(output.shape)

• torch.flatten: 将输入（Tensor）展平为一维张量

• batch_size 一般不展开，以MNIST数据集的一个 batch 为例将其依次转化为例：

  [64, 1, 28, 28] -> [64, 784] -> [64, 128]

神经网络-实践以及Sequential

import torch
from torch import nn
from torch.nn import Conv2d, MaxPool2d, Flatten, Linear


class Nn_SeqModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.conv1 = Conv2d(3, 32, 5, padding=2)
        self.maxpool1 = MaxPool2d(2)
        self.conv2 = Conv2d(32, 32, 5, padding=2)
        self.maxpool2 = MaxPool2d(2)
        self.conv3 = Conv2d(32, 64, 5, padding=2)
        self.maxpool3 = MaxPool2d(2)
        self.flatten = Flatten()
        self.linear1 = Linear(1024, 64)
        self.linear2 = Linear(64, 10)

    def forward(self, x):
        x = self.conv1(x)
        x = self.maxpool1(x)
        x = self.conv2(x)
        x = self.maxpool2(x)
        x = self.conv3(x)
        x = self.maxpool2(x)
        x = self.flatten(x)
        x = self.linear1(x)
        x = self.linear2(x)
        return x


if __name__ == '__main__':
    nn_seqmodel = Nn_SeqModel()
    print(nn_seqmodel)
    # 对网络模型进行检验
    input = torch.ones((64, 3, 32, 32))
    output = nn_seqmodel(input)
    print(output.shape)

Nn_SeqModel(
(conv1): Conv2d(3, 32, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))
(maxpool1): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(conv2): Conv2d(32, 32, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))
(maxpool2): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(conv3): Conv2d(32, 64, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))
(maxpool3): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(flatten): Flatten(start_dim=1, end_dim=-1)
(linear1): Linear(in_features=1024, out_features=64, bias=True)
(linear2): Linear(in_features=64, out_features=10, bias=True)
)
torch.Size([64, 10])

• Sequential 使代码更加简洁

import torch
from torch import nn
from torch.nn import Conv2d, MaxPool2d, Flatten, Linear, Sequential


class Nn_SeqModel(nn.Module):
    def __init__(self):
        super().__init__()
        # self.conv1 = Conv2d(3, 32, 5, padding=2)
        # self.maxpool1 = MaxPool2d(2)
        # self.conv2 = Conv2d(32, 32, 5, padding=2)
        # self.maxpool2 = MaxPool2d(2)
        # self.conv3 = Conv2d(32, 64, 5, padding=2)
        # self.maxpool3 = MaxPool2d(2)
        # self.flatten = Flatten()
        # self.linear1 = Linear(1024, 64)
        # self.linear2 = Linear(64, 10)
        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.conv1(x)
        # x = self.maxpool1(x)
        # x = self.conv2(x)
        # x = self.maxpool2(x)
        # x = self.conv3(x)
        # x = self.maxpool2(x)
        # x = self.flatten(x)
        # x = self.linear1(x)
        # x = self.linear2(x)
        x = self.model1(x)
        return x


if __name__ == '__main__':
    nn_seqmodel = Nn_SeqModel()
    print(nn_seqmodel)
    # 对网络模型进行检验
    input = torch.ones((64, 3, 32, 32))
    output = nn_seqmodel(input)
    print(output.shape)

Nn_SeqModel(
(model1): Sequential(
(0): Conv2d(3, 32, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))
(1): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(2): Conv2d(32, 32, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))
(3): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(4): Conv2d(32, 64, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))
(5): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(6): Flatten(start_dim=1, end_dim=-1)
(7): Linear(in_features=1024, out_features=64, bias=True)
(8): Linear(in_features=64, out_features=10, bias=True)
)
)
torch.Size([64, 10])

if __name__ == '__main__':
    nn_seqmodel = Nn_SeqModel()
    print(nn_seqmodel)
    # 对网络模型进行检验
    input = torch.ones((64, 3, 32, 32))
    output = nn_seqmodel(input)
    print(output.shape)
    # 查看网络结构
    writer = SummaryWriter('./logs_seq')
    writer.add_graph(nn_seqmodel, input)
    writer.close()

• 查看网络结构

损失函数与反向传播

• loos 损失函数

• 注意输入和输出

import torch
from torch.nn import L1Loss

inputs = torch.tensor([1, 2, 3], dtype=torch.float32)
targets = torch.tensor([1, 2, 5], dtype=torch.float32)

inputs = torch.reshape(inputs, (1, 1, 1, 3))
targets = torch.reshape(targets, (1, 1, 1, 3))

loss = L1Loss()  # reduction='sum'
result = loss(inputs, targets)
print(result)

tensor(0.6667)

• 交叉熵loss

x = torch.tensor([0.1, 0.2, 0.3])
y = torch.tensor([1])
x = torch.reshape(x, (1, 3))
loss_cross = nn.CrossEntropyLoss()
result_cross = loss_cross(x, y)
print(f"The result_cross of CrossEntropyLoss: {result_cross}")

The result_cross of CrossEntropyLoss: 1.1019428968429565

• 测试

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

dataset = torchvision.datasets.CIFAR10("./dataset", train=False, transform=torchvision.transforms.ToTensor(),
                                       download=True)
dataloader = DataLoader(dataset, batch_size=1)


class Nn_LossNetworkModel(nn.Module):
    def __init__(self):
        super().__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()

if __name__ == '__main__':
    nn_lossmodel = Nn_LossNetworkModel()
    for data in dataloader:
        imgs, targets = data
        outputs = nn_lossmodel(imgs)
        result_loss = loss(outputs, targets)
        print(f"the result_loss is : {result_loss}")

• 梯度下降 进行反向传播
• debug测试查看 grad

if __name__ == '__main__':
    nn_lossmodel = Nn_LossNetworkModel()
    for data in dataloader:
        imgs, targets = data
        outputs = nn_lossmodel(imgs)
        result_loss = loss(outputs, targets)
        # print(f"the result_loss is : {result_loss}")
        result_loss.backward()
        print("ok")

优化器

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

# 加载数据集转换为tensor类型
dataset = torchvision.datasets.CIFAR10("./dataset", train=False, transform=torchvision.transforms.ToTensor(),
                                       download=True)
# 使用DataLoader将数据集进行加载
dataloader = DataLoader(dataset, batch_size=1)


# 创建网络
class Nn_LossNetworkModel(nn.Module):
    def __init__(self):
        super().__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


if __name__ == '__main__':
    loss = nn.CrossEntropyLoss()
    nn_lossmodel = Nn_LossNetworkModel()
    optim = torch.optim.SGD(nn_lossmodel.parameters(), lr=0.01)
    for data in dataloader:
        imgs, targets = data
        outputs = nn_lossmodel(imgs)
        result_loss = loss(outputs, targets)
        optim.zero_grad()
        result_loss.backward()
        optim.step()

if __name__ == '__main__':
    loss = nn.CrossEntropyLoss()
    nn_lossmodel = Nn_LossNetworkModel()
    optim = torch.optim.SGD(nn_lossmodel.parameters(), lr=0.01)
    for epoch in range(20):
        running_loss = 0.0
        for data in dataloader:
            imgs, targets = data
            outputs = nn_lossmodel(imgs)
            result_loss = loss(outputs, targets)
            optim.zero_grad()
            result_loss.backward()
            optim.step()
            running_loss = running_loss + result_loss
        print("running_loss: ", running_loss)

Files already downloaded and verified
running_loss: tensor(18788.4355, grad_fn=)
running_loss: tensor(16221.9961, grad_fn=)

........

现有网络模型的使用以及修改

import torchvision
import torch
from torch import nn

# train_data = torchvision.datasets.ImageNet("./data_image_net", split="train",
#                                            transform=torchvision.transforms.ToTensor(), download=True)

vgg16_false = torchvision.models.vgg16(pretrained=False)
vgg16_true = torchvision.models.vgg16(pretrained=True)
print('ok')

print(vgg16_true)
train_data = torchvision.datasets.CIFAR10('./dataset', train=True, transform=torchvision.transforms.ToTensor(),
                                          download=True)
# vgg16_true.add_module('add_linear', nn.Linear(1000, 10))
vgg16_true.classifier.add_module('add_linear', nn.Linear(1000, 10))
print(vgg16_true)

print(vgg16_false)
vgg16_false.classifier[6] = nn.Linear(4096, 10)
print(vgg16_false)

网络模型的保存与读取

• save

import torch
import torchvision
from torch import nn

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")


# 陷阱
class Nn_Model(nn.Module):
    def __init__(self):
        super().__init__()
        self.conv1 = nn.Conv2d(3, 64, 3)

    def forward(self, x):
        x = self.conv1(x)
        return x


nn_model = Nn_Model()
torch.save(nn_model, "nnModel_method1.pth")

• load

import torch
import torchvision
from torch import nn
from p19_model_save import *

# 加载方式1 ---> 对应保存方式1 ，加载模型
model = torch.load("vgg16_method1.pth")
# print(model)

# 加载方式2
model2 = torch.load("vgg16_method2.pth")
print(model2)

# 方式2 的回复网络模型结构
vgg16 = torchvision.models.vgg16(pretrained=False)
vgg16.load_state_dict(torch.load("vgg16_method2.pth"))
print(vgg16)

# 陷阱1
# class Nn_Model(nn.Module):
#     def __init__(self):
#         super().__init__()
#         self.conv1 = nn.Conv2d(3, 64, 3)
#
#     def forward(self, x):
#         x = self.conv1(x)
#         return x


model1 = torch.load("nnModel_method1.pth")
print(model1)

完成的模型训练套路（一）

• 建包 train.py 和 model.py
• model.py

import torch
from torch import nn


# 搭建神经网络
class Nn_Neural_NetWork(nn.Module):
    def __init__(self):
        super().__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


if __name__ == '__main__':
    # 测试一下模型准确性
    nn_model = Nn_Neural_NetWork()
    input = torch.ones((64, 3, 32, 32))
    output = nn_model(input)
    print(output.shape)

• train.py

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

from model import *

# 准备数据集
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)
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_loader = DataLoader(train_data, batch_size=64)
test_loader = DataLoader(test_data, batch_size=64)

# 创建网络模型
nn_model = Nn_Neural_NetWork()

# 损失函数
loss_fn = nn.CrossEntropyLoss()

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

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

    # 训练步骤开始
    for data in train_loader:
        imgs, targets = data
        output = nn_model(imgs)
        loss = loss_fn(output, targets)

        # 优化器优化模型
        optimizer.zero_grad()
        loss.backward()
        optimizer.step()
        total_train_step += 1
        print("训练次数: {}, Loss: {}".format(total_train_step, loss.item()))

完成的模型训练套路（二）

• train.py
• 增加了tenorboard
• 增加了精确度Accuracy

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

from p20_model import *

# 准备数据集
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)
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_loader = DataLoader(train_data, batch_size=64)
test_loader = DataLoader(test_data, batch_size=64)

# 创建网络模型
nn_model = Nn_Neural_NetWork()

# 损失函数
loss_fn = nn.CrossEntropyLoss()

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

# （可加可不加） 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
    for data in train_loader:
        imgs, targets = data
        output = nn_model(imgs)
        loss = loss_fn(output, targets)

        # 优化器优化模型
        optimizer.zero_grad()
        loss.backward()
        optimizer.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
    total_test_loss = 0
    # 精确度
    total_accuracy = 0
    with torch.no_grad():
        for data in test_loader:
            imgs, targets = data
            outputs = nn_model(imgs)
            loss = loss_fn(outputs, targets)
            total_test_loss += loss
            accuracy = (outputs.argmax(1) == targets).sum()
            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Accuracy: {}".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 += 1

    # 保存模型结果
    torch.save(nn_model, "model_{}.pth".format(i))
    print("模型保存")

writer.close()

参考：
[1] bilibili-小土堆up