个人中心
gpt4 book ai didi

pytorch实现mnist分类的示例讲解

转载 作者：qq735679552 更新时间：2022-09-29 22:32:09 28 4
gpt4 key购买 nike

CFSDN坚持开源创造价值，我们致力于搭建一个资源共享平台，让每一个IT人在这里找到属于你的精彩世界.

这篇CFSDN的博客文章pytorch实现mnist分类的示例讲解由作者收集整理，如果你对这篇文章有兴趣，记得点赞哟.

torchvision包 包含了目前流行的数据集，模型结构和常用的图片转换工具.

torchvision.datasets中包含了以下数据集 。

MNIST COCO（用于图像标注和目标检测）(Captioning and Detection) LSUN Classification ImageFolder Imagenet-12 CIFAR10 and CIFAR100 STL10 。

torchvision.models 。

torchvision.models模块的 子模块中包含以下模型结构。 AlexNet VGG ResNet SqueezeNet DenseNet You can construct a model with random weights by calling its constructor

pytorch torchvision transform 。

对PIL.Image进行变换 。

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
from __future__ import print_function
import argparse #Python 命令行解析工具
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from torchvision import datasets, transforms
 
class Net(nn.Module):
   def __init__( self ):
     super (Net, self ).__init__()
     self .conv1 = nn.Conv2d( 1 , 10 , kernel_size = 5 )
     self .conv2 = nn.Conv2d( 10 , 20 , kernel_size = 5 )
     self .conv2_drop = nn.Dropout2d()
     self .fc1 = nn.Linear( 320 , 50 )
     self .fc2 = nn.Linear( 50 , 10 )
 
   def forward( self , x):
     x = F.relu(F.max_pool2d( self .conv1(x), 2 ))
     x = F.relu(F.max_pool2d( self .conv2_drop( self .conv2(x)), 2 ))
     x = x.view( - 1 , 320 )
     x = F.relu( self .fc1(x))
     x = F.dropout(x, training = self .training)
     x = self .fc2(x)
     return F.log_softmax(x, dim = 1 )
 
def train(args, model, device, train_loader, optimizer, epoch):
   model.train()
   for batch_idx, (data, target) in enumerate (train_loader):
     data, target = data.to(device), target.to(device)
     optimizer.zero_grad()
     output = model(data)
     loss = F.nll_loss(output, target)
     loss.backward()
     optimizer.step()
     if batch_idx % args.log_interval = = 0 :
       print ( 'Train Epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}' . format (
         epoch, batch_idx * len (data), len (train_loader.dataset),
         100. * batch_idx / len (train_loader), loss.item()))
 
def test(args, model, device, test_loader):
   model. eval ()
   test_loss = 0
   correct = 0
   with torch.no_grad():
     for data, target in test_loader:
       data, target = data.to(device), target.to(device)
       output = model(data)
       test_loss + = F.nll_loss(output, target, size_average = False ).item() # sum up batch loss
       pred = output. max ( 1 , keepdim = True )[ 1 ] # get the index of the max log-probability
       correct + = pred.eq(target.view_as(pred)). sum ().item()
 
   test_loss / = len (test_loader.dataset)
   print ( '\nTest set: Average loss: {:.4f}, Accuracy: {}/{} ({:.0f}%)\n' . format (
     test_loss, correct, len (test_loader.dataset),
     100. * correct / len (test_loader.dataset)))
 
def main():
   # Training settings
   parser = argparse.ArgumentParser(description = 'PyTorch MNIST Example' )
   parser.add_argument( '--batch-size' , type = int , default = 64 , metavar = 'N' ,
             help = 'input batch size for training (default: 64)' )
   parser.add_argument( '--test-batch-size' , type = int , default = 1000 , metavar = 'N' ,
             help = 'input batch size for testing (default: 1000)' )
   parser.add_argument( '--epochs' , type = int , default = 10 , metavar = 'N' ,
             help = 'number of epochs to train (default: 10)' )
   parser.add_argument( '--lr' , type = float , default = 0.01 , metavar = 'LR' ,
             help = 'learning rate (default: 0.01)' )
   parser.add_argument( '--momentum' , type = float , default = 0.5 , metavar = 'M' ,
             help = 'SGD momentum (default: 0.5)' )
   parser.add_argument( '--no-cuda' , action = 'store_true' , default = False ,
             help = 'disables CUDA training' )
   parser.add_argument( '--seed' , type = int , default = 1 , metavar = 'S' ,
             help = 'random seed (default: 1)' )
   parser.add_argument( '--log-interval' , type = int , default = 10 , metavar = 'N' ,
             help = 'how many batches to wait before logging training status' )
   args = parser.parse_args()
   use_cuda = not args.no_cuda and torch.cuda.is_available()
 
   torch.manual_seed(args.seed)
 
   device = torch.device( "cuda" if use_cuda else "cpu" )
 
   kwargs = { 'num_workers' : 1 , 'pin_memory' : True } if use_cuda else {}
   train_loader = torch.utils.data.DataLoader(
     datasets.MNIST( '../data' , train = True , download = True ,
             transform = transforms.Compose([
               transforms.ToTensor(),
               transforms.Normalize(( 0.1307 ,), ( 0.3081 ,))
             ])),
     batch_size = args.batch_size, shuffle = True , * * kwargs)
   test_loader = torch.utils.data.DataLoader(
     datasets.MNIST( '../data' , train = False , transform = transforms.Compose([
               transforms.ToTensor(),
               transforms.Normalize(( 0.1307 ,), ( 0.3081 ,))
             ])),
     batch_size = args.test_batch_size, shuffle = True , * * kwargs)
 
 
   model = Net().to(device)
   optimizer = optim.SGD(model.parameters(), lr = args.lr, momentum = args.momentum)
 
   for epoch in range ( 1 , args.epochs + 1 ):
     train(args, model, device, train_loader, optimizer, epoch)
     test(args, model, device, test_loader)
 
 
if __name__ = = '__main__' :
   main()

以上这篇pytorch实现mnist分类的示例讲解就是小编分享给大家的全部内容了，希望能给大家一个参考，也希望大家多多支持我.

原文链接：https://blog.csdn.net/KyrieHe/article/details/80516737 。

最后此篇关于pytorch实现mnist分类的示例讲解的文章就讲到这里了,如果你想了解更多关于pytorch实现mnist分类的示例讲解的内容请搜索CFSDN的文章或继续浏览相关文章，希望大家以后支持我的博客！ 。

28 4 0
文章推荐： pytorch GAN伪造手写体mnist数据集方式
文章推荐： pytorch 准备、训练和测试自己的图片数据的方法
文章推荐： Pytorch使用MNIST数据集实现CGAN和生成指定的数字方式
文章推荐： Pytorch使用MNIST数据集实现基础GAN和DCGAN详解
qq735679552
个人简介

我是一名优秀的程序员,十分优秀！

滴滴打车优惠券免费领取
滴滴打车优惠券
Copyright 2021 - 2024 cfsdn All Rights Reserved 蜀ICP备2022000587号
广告合作:1813099741@qq.com 6ren.com