10.10 学习笔记（搭建神经网络）上篇

搭建神经网络

CIFAR10 Model

---

推算公式

损失函数

1. 计算实际输出与目标函数之间的差距
2. 为更新输出提供一定的依据（反向传播，比如梯度）

代码

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


dataset = torchvision.datasets.CIFAR10(root="../dataset", train=False, transform=torchvision.transforms.ToTensor(), download=False)
dataloader = DataLoader(dataset, batch_size=64)
class MyCNN(nn.Module):
    def __init__(self):
        super(MyCNN, self).__init__()
        self.conv = Sequential(
            Conv2d(3, 32, kernel_size=5, padding=2),
            MaxPool2d(2),
            Conv2d(32, 32, kernel_size=5, padding=2),
            MaxPool2d(2),
            Conv2d(32, 64, kernel_size=5, padding=2),
            MaxPool2d(2),
            Flatten(),
            Linear(1024, 64),
            Linear(64, 10)
        )

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


mycnn = MyCNN()

# 损失函数 & 优化器
loss = nn.CrossEntropyLoss()
optim = torch.optim.SGD(mycnn.parameters(), lr=0.01)

for epoch in range(20):
    running_loss = 0
    for data in dataloader:
        imgs, target = data
        ouput = mycnn(imgs)
        result_loss = loss(ouput, target)
        optim.zero_grad()
        result_loss.backward()
        optim.step()
        running_loss += result_loss
    print(running_loss)

# # 用来验证模型
# input = torch.ones((64, 3, 32, 32))
# output = mycnn(input)
# print(output.shape)
# # 计算图
# writer = SummaryWriter("torch_seq")
# writer.add_graph(mycnn, input)
# writer.close()

计算图

模型的加载与保存

代码给出说明（保存方式与加载方式需要一一对应）

import os

import torch
import torchvision

from torch import nn
from torch.nn import Linear
from torchvision.models import VGG16_Weights
os.environ['TORCH_HOME'] = './torch_cache'

vgg16_true = torchvision.models.vgg16(weights=VGG16_Weights.IMAGENET1K_V1)
train_data = torchvision.datasets.CIFAR10(root="../dataset", train=True, transform=torchvision.transforms.ToTensor(), download=True)
vgg16_true.classifier.add_module('add_Linear', Linear(1000, 10))

vgg16_false = torchvision.models.vgg16(weights=VGG16_Weights.DEFAULT)

# print(vgg16_false)

# print(vgg16_true)

# 保存方式1 模型结构+模型参数
torch.save(vgg16_false, "vgg_method1.pth")

# 保存方式2 模型参数（官方推荐）
torch.save(vgg16_false.state_dict(), "vgg_method2.pth")

# 加载方式1
model = torch.load("vgg_method1.pth")

# 加载方式2
model2 = torchvision.models.vgg16(weights=VGG16_Weights.DEFAULT)
model2.state_dict(torch.load("vgg_method2.pth"))