目录
前言
B战UP主我是土堆的pytorch教程，记录下自己学习的笔记，仅供自己参考
作者：我是土堆
环境：python-3.6.13 pytorch-1.10.1
1.Python两大法宝函数(Pytorch)
-
可以将pytoch看成一个工具箱
-
dir()函数，能让我们知道工具箱以及工具箱中的分隔区有什么东西。
- 当看到dir()输出内容为__xxx__时，代表不可修改为一个函数，此时可使用help()来查看其使用方法
-
help()函数，能让我们知道每个工具是如何使用的，工具的使用方法
import torch
dir(torch)
dir(torch.cuda)
dir(torch.cuda.is_available())
help(torch.cuda.is_available)
2.数据加载
两个类
- Dataset 提供一种方式去获取数据及其label
- 如何获取每一个数据及其label
- 告诉我们总共有多少的数据
- Dataloader 为后面的网络提供不同的数据形式
示例代码如下：
from torch.utils.data import Dataset
from PIL import Image
import os
class MyData(Dataset):
def __init__(self, root_dir, label_dir) -> None:
super().__init__()
self.root_dir = root_dir
self.label_dir = label_dir
self.path = os.path.join(self.root_dir, self.label_dir)
self.img_path = os.listdir(self.path)
def __getitem__(self, idx):
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):
return len(self.img_path)
if __name__ == "__main__":
root_dir = "hymenoptera_data\\train"
ants_label_dir = "ants"
ants_dataset = MyData(root_dir, ants_label_dir)
bees_label_dir = "bees"
bees_dataset = MyData(root_dir, ants_label_dir)
train_dataset = ants_dataset + bees_dataset
img, label = train_dataset[244] # 调用__getitem__方法
img.show()
3.TensorBoard
tensorboard是一个可视化工具，在模型训练时经常使用，博主不太关注，请各位自行观看相关视频讲解
writer.add_scalar()
- global_step代表x轴
- scalar_value代表y轴
tensorboard启动指令如下：
tensorboard --logdir=logs --port=6007
示例代码如下：
from torch.utils.tensorboard import SummaryWriter
from PIL import Image
import numpy as np
writer = SummaryWriter("logs")
image_path = "data/train/bees_image/16838648_415acd9e3f.jpg"
img_PIL = Image.open(image_path)
img_array = np.array(img_PIL) # HWC
writer.add_image("test", img_array, 2, dataformats="HWC")
# y = 2x
for i in range(100):
writer.add_scalar("y=2x", 2*i, i)
writer.close()
4.torchvison中的transforms
transforms主要是指图片中的一些常见变换
transforms的结构和用法
-
结构
-
pycharm中的structure(Alt+7)可以查看package的目录结构
-
transforms.py是工具箱，包含totensor、resize等工具模板
-
图片->transform工具箱中工具模板->创建具体的工具->结果
-
-
用法
示例代码如下:
from torchvision import transforms
from PIL import Image
from torch.utils.tensorboard import SummaryWriter
# python的用法->tensor数据类型
# 通过 transforms.ToTensor去看两个问题
# 1. transform该如何使用(python)
# 2. 为什么我们需要Tensor数据类型
# tensor数据类型包装了我们神经网络所需要的一些参数,比如backward,grad,device...
# 图片->transform工具箱中的工具模板->创建具体的工具->结果
img_path = "data/train/ants_image/0013035.jpg"
# 1.图片
img = Image.open(img_path)
writer = SummaryWriter("logs")
# 2.使用transform工具箱中的工具模板，创建具体的工具
tensor_trans = transforms.ToTensor()
# 3.结果
tensor_img = tensor_trans(img)
print(tensor_img)
writer.add_image("Tensor_img", tensor_img)
writer.close()
5.常用transform
python中__call__用法
- 直接使用对象加上()进行操作
- pycharm中按住ctrl+p查看需要哪些参数
示例代码如下:
class Person:
def __call__(self, name):
print("__call__" + " Hello " + name)
def hello(self, name):
print("Hello " + name)
person = Person()
person("Tom")
person.hello("Jerry")
transform的使用
-
关注输入和输出类型
-
多看官方文档
-
关注方法需要什么参数
示例代码如下:
from PIL import Image
from torchvision import transforms
from torch.utils.tensorboard import SummaryWriter
writer = SummaryWriter("logs")
img = Image.open("images/pytorch.png")
print(img.size)
# 1. ToTensor
trans_totensor = transforms.ToTensor()
img_tensor = trans_totensor(img)
writer.add_image("ToTensor", img_tensor)
# 2. Normalize
# output[channel] = (input[channel] - mean[channel]) / std[channel]
print(img_tensor[0][0][0])
trans_norm = transforms.Normalize([0.5, 0.5, 0.5], [0.5, 0.5, 0.5])
img_norm = trans_norm(img_tensor)
print(img_norm[0][0][0])
writer.add_image("Normalize", img_norm)
# 3. Resize
trans_resize = transforms.Resize((640, 640))
# img PIL -> resize -> img_resize PIL
img_resize = trans_resize(img)
# img_resize PIL -> totensor -> img_resize tensor
img_resize = trans_totensor(img_resize)
writer.add_image("Resize", img_resize)
# 4. Compose
trans_resize_2 = transforms.Resize((640, 640))
trans_compose = transforms.Compose([trans_resize_2, trans_totensor]) # list列表
img_resize_2 = trans_compose(img)
writer.add_image("Compose", img_resize_2)
# 5. RandomCrop
trans_random = transforms.RandomCrop(256)
trans_compose_2 = transforms.Compose([trans_random, trans_totensor])
# for i in range(10):
img_crop = trans_compose_2(img)
writer.add_image("RandomCrop", img_crop)
writer.close()
不知道返回值的时候可尝试以下几种方法：
- print(type())
- debug
6.torchvision中的数据集使用
torchvision数据集官方文档参考
-
对于数据集的下载tips:可将下载链接复制到迅雷中进行下载
示例代码1:
- 获取CIFAR10数据集
import torchvision
dataset_transform = torchvision.transforms.Compose([
torchvision.transforms.ToTensor()
])
train_set = torchvision.datasets.CIFAR10(root="./dataset", train=True, download=True)
test_set = torchvision.datasets.CIFAR10(root="./dataset", train=False, 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()
示例代码2:
- 针对数据集进行transform
import torchvision
from torch.utils.tensorboard import SummaryWriter
dataset_transform = torchvision.transforms.Compose([
torchvision.transforms.ToTensor()
])
train_set = torchvision.datasets.CIFAR10(root="./dataset", train=True, transform=dataset_transform, download=True)
test_set = torchvision.datasets.CIFAR10(root="./dataset", train=False, transform=dataset_transform, download=True)
print(test_set[0])
# tensorboard显示tensor类型的图片
writer = SummaryWriter("logs")
for i in range(10):
img, target = test_set[i]
writer.add_image("test_set", img, i)
writer.close()
7.DataLoader的使用
torch中DataLoader官方文档
示例代码如下:
import torchvision
from torch.utils.data import DataLoader
from torch.utils.tensorboard import SummaryWriter
test_data = torchvision.datasets.CIFAR10(root="./dataset", train=False, transform=torchvision.transforms.ToTensor())
test_loader = DataLoader(dataset=test_data, batch_size=64, shuffle=True, num_workers=0, drop_last=False)
# 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("logs")
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()
8.神经网络的基本骨架-nn.Module的使用
torch.nn.Module官方文档
示例代码如下:
import torch
from torch import nn
class Model(nn.Module):
def __init__(self) -> None:
super().__init__()
def forward(self, input):
output = input + 1
return output
model = Model()
x = torch.tensor(1.0)
output = model(x)
print(output)
9.神经网络
torch.nn是torch.nn.functional的封装,可以看成是包含的关系
9.1.卷积层
torch.nn.Conv2d官方文档
-
nn.Conv2d中参数dilation意指空洞卷积
示例代码1:
- conv中的stride和padding
import torch
import torch.nn.functional as F
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]
])
kernel = torch.tensor([
[1, 2, 1],
[0, 1, 0],
[2, 1, 0]
])
# torch提供的尺寸变换
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)
output3 = F.conv2d(input, kernel, stride=1, padding=1)
print(output3)
示例代码2：
- torch.nn.Conv2d使用
import torch
import torchvision
from torch import nn
from torch.utils.data import DataLoader
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 Model(nn.Module):
def __init__(self) -> None:
super().__init__()
self.conv1 = nn.Conv2d(in_channels=3, out_channels=3, kernel_size=3, stride=1, padding=0)
def forward(self, x):
x = self.conv1(x)
return x
model = Model()
writer = SummaryWriter("./logs")
step = 0
for data in dataloader:
imgs, targets = data
output = model(imgs)
# print(imgs.shape)
# print(output.shape)
# torch.Size([64, 3, 32, 32])
writer.add_images("input", imgs, step)
# torch.Size([64, 3, 30, 30])
# output = torch.reshape(output, (-1, 3, 30, 30))
writer.add_images("output", output, step)
step = step + 1
9.2.最大池化的使用
torch.nn.MaxPool2d官方文档
示例代码1：
import torch
import torchvision
import torch.nn as nn
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]], dtype=torch.float32)
input = torch.reshape(input, (-1, 1, 5, 5))
print(input.shape)
class Model(nn.Module):
def __init__(self) -> None:
super().__init__()
self.maxpool1 = nn.MaxPool2d(kernel_size=3, ceil_mode=True)
def forward(self, x):
output = self.maxpool1(input)
return output
model = Model()
output = model(input)
print(output)
示例代码2：
import torch
import torchvision
import torch.nn as nn
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)
class Model(nn.Module):
def __init__(self) -> None:
super().__init__()
self.maxpool1 = nn.MaxPool2d(kernel_size=3, ceil_mode=True)
def forward(self, x):
output = self.maxpool1(x)
return output
model = Model()
writer = SummaryWriter("./logs")
step = 0
for data in dataloader:
imgs, targets = data
output = model(imgs)
writer.add_images("imgs", imgs, step)
writer.add_images("output", output, step)
step = step + 1
writer.close()
9.3.非线性激活
Non-linear Activations官方文档
ReLU中inplace参数，是否对原变量进行替换
示例代码1：
import torch
import torch.nn as nn
input = torch.tensor([[1, -0.5],
[-1, 3]])
input = torch.reshape(input, (-1, 1, 2, 2))
print(input.shape)
class Model(nn.Module):
def __init__(self) -> None:
super().__init__()
self.relu1 = nn.ReLU()
def forward(self, input):
output = self.relu1(input)
return output
model = Model()
output = model(input)
print(output)
示例代码2：
import torch
import torch.nn as nn
import torchvision
from torch.utils.data import DataLoader
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 Model(nn.Module):
def __init__(self) -> None:
super().__init__()
self.sigmoid = nn.Sigmoid()
def forward(self, input):
output = self.sigmoid(input)
return output
model = Model()
writer = SummaryWriter("./logs")
step = 0
for data in dataloader:
imgs, targets = data
output = model(imgs)
writer.add_images("imgs", imgs, step)
writer.add_images("output", output, step)
step = step + 1
writer.close()
9.4.线性层及其他层介绍
其他层官方文档
示例代码1：
- Linear Layer
import torch
import torchvision
import torch.nn as nn
from torch.utils.data import DataLoader
dataset = torchvision.datasets.CIFAR10("./dataset", train=False, transform=torchvision.transforms.ToTensor(), download=True)
dataloader = DataLoader(dataset, batch_size=64)
class Model(nn.Module):
def __init__(self) -> None:
super().__init__()
self.linear1 = nn.Linear(196608, 10)
def forward(self, input):
output = self.linear1(input)
return output
model = Model()
for data in dataloader:
imgs, targets = data
print(imgs.shape)
# output = torch.reshape(imgs, (1, 1, 1, -1))
output = torch.flatten(imgs)
print(output.shape)
output = model(output)
print(output.shape)
9.5.搭建小实战和Sequential的使用
torch.nn.Sequential官方文档
如何检查搭建的网络正确性
- 使用假想的输入即dummy_input查看对应的输出是否正确
- 使用tensorboard查看网络结构
示例代码：
import torch
import torch.nn as nn
from torch.utils.tensorboard import SummaryWriter
# CIFAR 10 Model
class Model(nn.Module):
def __init__(self) -> None:
super().__init__()
self.conv1 = nn.Conv2d(3, 32, 5, 1, 2)
self.maxpool1 = nn.MaxPool2d(2)
self.conv2 = nn.Conv2d(32, 32, 5, 1, 2)
self.maxpool2 = nn.MaxPool2d(2)
self.conv3 = nn.Conv2d(32, 64, 5, 1, 2)
self.maxpool3 = nn.MaxPool2d(2)
self.flatten = nn.Flatten()
self.linear1 = nn.Linear(1024, 64)
self.linear2 = nn.Linear(64, 10)
self.model1 = 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(1024, 64),
nn.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.maxpool3(x)
# x = self.flatten(x)
# x = self.linear1(x)
# x = self.linear2(x)
x = self.model1(x)
return x
model = Model()
print(model)
dummy_input = torch.ones((64, 3, 32, 32))
output = model(dummy_input)
print(output.shape)
writer = SummaryWriter("./logs")
writer.add_graph(model, dummy_input)
writer.close()
10.损失函数与反向传播
损失函数
-
计算实际输出和目标之间的差距
-
为我们更新输出提供一定的依据(反向传播)
示例代码1：
- 各种损失函数
from unittest import result
import torch
from torch.nn import L1Loss, MSELoss, CrossEntropyLoss
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))
# L1Losss
loss = L1Loss(reduction='sum') # sum mean
result = loss(inputs, targets)
# MSELoss
loss_mse = MSELoss()
result_mse = loss_mse(inputs, targets)
# CrossEntropyLoss
x = torch.tensor([0.1, 0.2, 0.3])
y = torch.tensor([1])
x = torch.reshape(x, (1, 3))
loss_cross = CrossEntropyLoss()
result_cross = loss_cross(x, y)
print(result_cross)
示例代码2：
- 反向传播检查梯度
import torch
import torch.nn as nn
from torch.utils.tensorboard import SummaryWriter
import torchvision
from torch.utils.data import DataLoader
dataset = torchvision.datasets.CIFAR10("./dataset", train=False, download=True,
transform=torchvision.transforms.ToTensor())
dataloader = DataLoader(dataset, batch_size=1)
# CIFAR 10 Model
class Model(nn.Module):
def __init__(self) -> None:
super().__init__()
self.model1 = 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(1024, 64),
nn.Linear(64, 10)
)
def forward(self, x):
x = self.model1(x)
return x
model = Model()
loss = nn.CrossEntropyLoss()
for data in dataloader:
imgs, targets = data
outputs = model(imgs)
result_loss = loss(outputs, targets)
# result_loss.backward()
11.优化器
torch.optim官方文档
import torch
import torch.nn as nn
from torch.utils.tensorboard import SummaryWriter
import torchvision
from torch.utils.data import DataLoader
dataset = torchvision.datasets.CIFAR10("./dataset", train=False, download=True,
transform=torchvision.transforms.ToTensor())
dataloader = DataLoader(dataset, batch_size=1)
# CIFAR 10 Model
class Model(nn.Module):
def __init__(self) -> None:
super().__init__()
self.model1 = 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(1024, 64),
nn.Linear(64, 10)
)
def forward(self, x):
x = self.model1(x)
return x
model = Model()
loss = nn.CrossEntropyLoss()
optim = torch.optim.SGD(model.parameters(), lr=0.01)
for epoch in range(20):
running_loss = 0.0
for data in dataloader:
imgs, targets = data
outputs = model(imgs)
result_loss = loss(outputs, targets)
optim.zero_grad()
result_loss.backward() # 节点梯度
optim.step() # 更新weight
running_loss = running_loss + result_loss
print(running_loss)
12.现有网络模型的使用及其修改
torchvision.model官方文档
示例代码：
import torchvision
from torch.utils.data import DataLoader
import torch.nn as nn
# train_data = torchvision.datasets.ImageNet("./ImageNet", split='train', download=True, transform=torchvision.transforms.ToTensor())
vgg16_false = torchvision.models.vgg16(pretrained=False)
# vgg16_True = torchvision.models.vgg16(pretrained=True)
dataset = torchvision.datasets.CIFAR10("./dataset", train=False, download=True,
transform=torchvision.transforms.ToTensor())
# 修改网络-方法1.新增层
# vgg16_false.add_module('add_linear', nn.Linear(1000, 10))
vgg16_false.classifier.add_module('add_linear', nn.Linear(1000, 10))
# 修改网络-方法2.修改原有层
vgg16_false.classifier[6] = nn.Linear(4096, 10)
print(vgg16_false)
13.网络模型的保存与读取
示例代码1：
- 模型的保存
from turtle import forward
import torchvision
import torch
import torch.nn as nn
vgg16 = torchvision.models.vgg16(pretrained=False)
# 保存方法1-模型结构+模型参数
torch.save(vgg16, "vgg16_model1.pth")
# 保存方法2-模型参数(官方推荐)
torch.save(vgg16.state_dict(), "vgg16_model2.pth")
# 陷阱
class Model(nn.Module):
def __init__(self) -> None:
super().__init__()
self.conv1 = nn.Conv2d(3, 64, kernel_size=3)
def forward(self, x):
x = self.conv1(x)
return x
model = Model()
torch.save(model, "model.pth")
示例代码2：
- 模型的读取
import torch
import torchvision
import torch.nn as nn
# 方式1->保存方式1,加载模型
model = torch.load("vgg16_model1.pth")
# print(model)
# 方式2->加载模型
model = torch.load("vgg16_model2.pth")
# 如何从网络参数恢复到网络模型
vgg16 = torchvision.models.vgg16(pretrained=False)
vgg16.load_state_dict(torch.load("vgg16_model2.pth"))
print(vgg16)
# 陷阱-需要将网络定义放在此处
class Model(nn.Module):
def __init__(self) -> None:
super().__init__()
self.conv1 = nn.Conv2d(3, 64, kernel_size=3)
def forward(self, x):
x = self.conv1(x)
return x
model = torch.load("model.pth")
print(model)
14.完整的模型训练套路
示例代码1：
- 完整训练流程
import torch
import torchvision
from torch.utils.data import DataLoader
import torch.nn as nn
from model import *
train_data = torchvision.datasets.CIFAR10("./dataset", train=True, transform=torchvision.transforms.ToTensor(), download=True)
test_data = torchvision.datasets.CIFAR10("./dataset", train=False, transform=torchvision.transforms.ToTensor(), download=True)
# 数据集大小
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
model = Model()
# 损失函数
loss_fn = nn.CrossEntropyLoss()
# 优化器
learnig_rate = 0.01
# learnig_rate = 1e-2
optimizer = torch.optim.SGD(model.parameters(), lr=learnig_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_dataloader:
imgs, targets = data
outputs = model(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
# item--将tensor数据类型转换为数字
print("训练次数: {}, loss: {}".format(total_train_step, loss.item()))
示例代码2：
- 添加测试步骤及tensorboard可视化
import torch
import torchvision
from torch.utils.data import DataLoader
import torch.nn as nn
from model import *
from torch.utils.tensorboard import SummaryWriter
train_data = torchvision.datasets.CIFAR10("./dataset", train=True, transform=torchvision.transforms.ToTensor(), download=True)
test_data = torchvision.datasets.CIFAR10("./dataset", train=False, transform=torchvision.transforms.ToTensor(), download=True)
# 数据集大小
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
model = Model()
# 损失函数
loss_fn = nn.CrossEntropyLoss()
# 优化器
learnig_rate = 0.01
# learnig_rate = 1e-2
optimizer = torch.optim.SGD(model.parameters(), lr=learnig_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")
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_dataloader:
imgs, targets = data
outputs = model(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
# item--将tensor数据类型转换为数字
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
total_accuracy = 0
with torch.no_grad():
for data in test_dataloader:
imgs, targets = data
outputs = model(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 = total_test_step + 1
torch.save(model, "model_{}.pth".format(i))
print("模型已保存")
writer.close()
model.train model.eval的作用
-
如果网络模型中包含Dropout、BatchNorm等层时需要指定
15.利用GPU训练
示例代码1：
- 方法1：网络模型、数据(输入、标注)、损失函数 .cuda()
import torch
import torchvision
from torch.utils.data import DataLoader
import torch.nn as nn
from torch.utils.tensorboard import SummaryWriter
import time
train_data = torchvision.datasets.CIFAR10("./dataset", train=True, transform=torchvision.transforms.ToTensor(), download=True)
test_data = torchvision.datasets.CIFAR10("./dataset", train=False, transform=torchvision.transforms.ToTensor(), download=True)
# 数据集大小
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 Model(nn.Module):
def __init__(self) -> None:
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(1024, 64),
nn.Linear(64, 10)
)
def forward(self, x):
x = self.model(x)
return x
model = Model()
if torch.cuda.is_available():
model = model.cuda() # 网络模型->cuda
# 损失函数
loss_fn = nn.CrossEntropyLoss()
if torch.cuda.is_available():
loss_fn = loss_fn.cuda() # 损失函数->cuda
# 优化器
learnig_rate = 0.01
# learnig_rate = 1e-2
optimizer = torch.optim.SGD(model.parameters(), lr=learnig_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")
start_time = time.time()
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_dataloader:
imgs, targets = data
if torch.cuda.is_available():
imgs = imgs.cuda() # 数据->cuda
targets = targets.cuda()
outputs = model(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
# item--将tensor数据类型转换为数字
if total_train_step % 100 == 0:
end_time = time.time()
print(end_time - start_time)
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
total_accuracy = 0
with torch.no_grad():
for data in test_dataloader:
imgs, targets = data
if torch.cuda.is_available():
imgs = imgs.cuda()
targets = targets.cuda()
outputs = model(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 = total_test_step + 1
torch.save(model, "model_{}.pth".format(i))
# torch.save(model.state_dict(), "model_{}.pth".format(i))
print("模型已保存")
writer.close()
示例代码2：
- 方法2：.to(device) device = torch.device(“cpu”)
import torch
import torchvision
from torch.utils.data import DataLoader
import torch.nn as nn
from torch.utils.tensorboard import SummaryWriter
import time
# 定义训练的设备
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
train_data = torchvision.datasets.CIFAR10("./dataset", train=True, transform=torchvision.transforms.ToTensor(), download=True)
test_data = torchvision.datasets.CIFAR10("./dataset", train=False, transform=torchvision.transforms.ToTensor(), download=True)
# 数据集大小
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 Model(nn.Module):
def __init__(self) -> None:
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(1024, 64),
nn.Linear(64, 10)
)
def forward(self, x):
x = self.model(x)
return x
model = Model()
model.to(device) # 网络模型->cuda
# 损失函数
loss_fn = nn.CrossEntropyLoss()
loss_fn.to(device) # 损失函数->cuda
# 优化器
learnig_rate = 0.01
# learnig_rate = 1e-2
optimizer = torch.optim.SGD(model.parameters(), lr=learnig_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")
start_time = time.time()
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_dataloader:
imgs, targets = data
imgs = imgs.to(device) # 数据->cuda
targets = targets.to(device)
outputs = model(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
# item--将tensor数据类型转换为数字
if total_train_step % 100 == 0:
end_time = time.time()
# print(end_time - start_time)
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
total_accuracy = 0
with torch.no_grad():
for data in test_dataloader:
imgs, targets = data
imgs = imgs.to(device) # 数据->cuda
targets = targets.to(device)
outputs = model(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 = total_test_step + 1
torch.save(model, "model_{}.pth".format(i))
# torch.save(model.state_dict(), "model_{}.pth".format(i))
print("模型已保存")
writer.close()
16.完整的模型验证套路
示例代码：
from pydoc import importfile
import torch
from PIL import Image
import torchvision
import torch.nn as nn
image_path = "./images/airplane.jpg"
image = Image.open(image_path)
# print(image)
image = image.convert("RGB")
transform = torchvision.transforms.Compose([
torchvision.transforms.Resize([32,32]),
torchvision.transforms.ToTensor()
])
image = transform(image)
# print(image.shape)
class Model(nn.Module):
def __init__(self) -> None:
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(1024, 64),
nn.Linear(64, 10)
)
def forward(self, x):
x = self.model(x)
return x
model = torch.load("model_9.pth")
# model = torch.load("model_9.pth", map_location=torch.device("cpu"))
# print(model)
image = torch.reshape(image, (1, 3, 32, 32))
image = image.cuda()
model.eval()
with torch.no_grad():
output = model(image)
print(output)
print(output.argmax(1))
17.看看开源项目
- 开源项目参数解析中建议将required=True删除添加default=xxx
结语
挺不错的pytorch入门教程,感兴趣的可以学习下。
本篇博客仅记录自己学习过程的笔记,仅供自己参考。
287
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
