chap5 CNN

卷积神经网络（CNN）

问题描述：

利用卷积神经网络，实现对MNIST数据集的分类问题

数据集：

MNIST数据集包括60000张训练图片和10000张测试图片。图片样本的数量已经足够训练一个很复杂的模型（例如 CNN的深层神经网络）。它经常被用来作为一个新 的模式识别模型的测试用例。而且它也是一个方便学生和研究者们执行用例的数据集。除此之外，MNIST数据集是一个相对较小的数据集，可以在你的笔记本CPUs上面直接执行

题目要求

Pytorch版本的卷积神经网络需要补齐self.conv1中的nn.Conv2d()和self.conv2()的参数，还需要填写x=x.view()中的内容。
训练精度应该在96%以上。

import os
import torch
import torch.nn as nn
from torch.autograd import Variable
import torch.utils.data as Data
import torchvision
import torch.nn.functional as F
import numpy as np
learning_rate = 1e-4
keep_prob_rate = 0.7 #
max_epoch = 3
BATCH_SIZE = 50DOWNLOAD_MNIST = False
if not(os.path.exists('./mnist/')) or not os.listdir('./mnist/'):# not mnist dir or mnist is empyt dirDOWNLOAD_MNIST = Truetrain_data = torchvision.datasets.MNIST(root='./mnist/',train=True, transform=torchvision.transforms.ToTensor(), download=DOWNLOAD_MNIST,)
train_loader = Data.DataLoader(dataset = train_data ,batch_size= BATCH_SIZE ,shuffle= True)test_data = torchvision.datasets.MNIST(root = './mnist/',train = False)
test_x = Variable(torch.unsqueeze(test_data.test_data,dim  = 1),volatile = True).type(torch.FloatTensor)[:500]/255.
test_y = test_data.test_labels[:500].numpy()class CNN(nn.Module):def __init__(self):super(CNN, self).__init__()self.conv1 = nn.Sequential(nn.Conv2d( # ???# patch 7 * 7 ; 1  in channels ; 32 out channels ; ; stride is 1# padding style is same(that means the convolution opration's input and output have the same size)in_channels=1,out_channels=32,kernel_size=7,stride=1,padding=3,),nn.ReLU(),        # activation functionnn.MaxPool2d(2),  # pooling operation)self.conv2 = nn.Sequential( # ???# line 1 : convolution function, patch 5*5 , 32 in channels ;64 out channels; padding style is same; stride is 1# line 2 : choosing your activation funciont# line 3 : pooling operation function.nn.Conv2d(in_channels=32, out_channels=64, kernel_size=5, padding=2, stride=1),nn.ReLU(),nn.AvgPool2d(2),)self.out1 = nn.Linear( 7*7*64 , 1024 , bias= True)   # full connection layer oneself.dropout = nn.Dropout(keep_prob_rate)self.out2 = nn.Linear(1024,10,bias=True)def forward(self, x):x = self.conv1(x)x = self.conv2(x)x = x.view(-1, 7*7*64)  # flatten the output of coonv2 to (batch_size ,32 * 7 * 7)    # ???out1 = self.out1(x)out1 = F.relu(out1)out1 = self.dropout(out1)out2 = self.out2(out1)output = F.softmax(out2)return outputdef test(cnn):global predictiony_pre = cnn(test_x)_,pre_index= torch.max(y_pre,1)pre_index= pre_index.view(-1)prediction = pre_index.data.numpy()correct  = np.sum(prediction == test_y)return correct / 500.0def train(cnn):optimizer = torch.optim.Adam(cnn.parameters(), lr=learning_rate )loss_func = nn.CrossEntropyLoss()for epoch in range(max_epoch):for step, (x_, y_) in enumerate(train_loader):x ,y= Variable(x_),Variable(y_)output = cnn(x)  loss = loss_func(output,y)optimizer.zero_grad()loss.backward()optimizer.step()if step != 0 and step % 20 ==0:print("=" * 10,step,"="*5,"="*5, "test accuracy is ",test(cnn) ,"=" * 10 )if __name__ == '__main__':cnn = CNN()train(cnn)

训练结果为：

========== 20 ===== ===== test accuracy is  0.224 ==========
========== 40 ===== ===== test accuracy is  0.362 ==========
========== 60 ===== ===== test accuracy is  0.402 ==========
========== 80 ===== ===== test accuracy is  0.51 ==========
========== 100 ===== ===== test accuracy is  0.608 ==========
========== 120 ===== ===== test accuracy is  0.624 ==========
========== 140 ===== ===== test accuracy is  0.708 ==========
========== 160 ===== ===== test accuracy is  0.684 ==========
========== 180 ===== ===== test accuracy is  0.738 ==========
========== 200 ===== ===== test accuracy is  0.766 ==========
========== 220 ===== ===== test accuracy is  0.778 ==========
========== 240 ===== ===== test accuracy is  0.796 ==========
========== 260 ===== ===== test accuracy is  0.802 ==========
========== 280 ===== ===== test accuracy is  0.81 ==========
========== 300 ===== ===== test accuracy is  0.812 ==========
========== 320 ===== ===== test accuracy is  0.82 ==========
========== 340 ===== ===== test accuracy is  0.848 ==========
========== 360 ===== ===== test accuracy is  0.83 ==========
========== 380 ===== ===== test accuracy is  0.852 ==========
========== 400 ===== ===== test accuracy is  0.852 ==========
========== 420 ===== ===== test accuracy is  0.856 ==========
========== 440 ===== ===== test accuracy is  0.874 ==========
========== 460 ===== ===== test accuracy is  0.85 ==========
========== 480 ===== ===== test accuracy is  0.874 ==========
========== 500 ===== ===== test accuracy is  0.864 ==========
========== 520 ===== ===== test accuracy is  0.858 ==========
========== 540 ===== ===== test accuracy is  0.884 ==========
========== 560 ===== ===== test accuracy is  0.872 ==========
========== 580 ===== ===== test accuracy is  0.9 ==========
========== 600 ===== ===== test accuracy is  0.88 ==========
========== 620 ===== ===== test accuracy is  0.886 ==========
========== 640 ===== ===== test accuracy is  0.882 ==========
========== 660 ===== ===== test accuracy is  0.886 ==========
========== 680 ===== ===== test accuracy is  0.876 ==========
========== 700 ===== ===== test accuracy is  0.882 ==========
========== 720 ===== ===== test accuracy is  0.886 ==========
========== 740 ===== ===== test accuracy is  0.894 ==========
========== 760 ===== ===== test accuracy is  0.894 ==========
========== 780 ===== ===== test accuracy is  0.9 ==========
========== 800 ===== ===== test accuracy is  0.898 ==========
========== 820 ===== ===== test accuracy is  0.912 ==========
========== 840 ===== ===== test accuracy is  0.894 ==========
========== 860 ===== ===== test accuracy is  0.898 ==========
========== 880 ===== ===== test accuracy is  0.888 ==========
========== 900 ===== ===== test accuracy is  0.896 ==========
========== 920 ===== ===== test accuracy is  0.888 ==========
========== 940 ===== ===== test accuracy is  0.91 ==========
========== 960 ===== ===== test accuracy is  0.908 ==========
========== 980 ===== ===== test accuracy is  0.918 ==========
========== 1000 ===== ===== test accuracy is  0.906 ==========
========== 1020 ===== ===== test accuracy is  0.908 ==========
========== 1040 ===== ===== test accuracy is  0.906 ==========
========== 1060 ===== ===== test accuracy is  0.914 ==========
========== 1080 ===== ===== test accuracy is  0.908 ==========
========== 1100 ===== ===== test accuracy is  0.906 ==========
========== 1120 ===== ===== test accuracy is  0.906 ==========
========== 1140 ===== ===== test accuracy is  0.924 ==========
========== 1160 ===== ===== test accuracy is  0.918 ==========
========== 1180 ===== ===== test accuracy is  0.904 ==========
========== 20 ===== ===== test accuracy is  0.924 ==========
========== 40 ===== ===== test accuracy is  0.908 ==========
========== 60 ===== ===== test accuracy is  0.92 ==========
========== 80 ===== ===== test accuracy is  0.91 ==========
========== 100 ===== ===== test accuracy is  0.926 ==========
========== 120 ===== ===== test accuracy is  0.91 ==========
========== 140 ===== ===== test accuracy is  0.922 ==========
========== 160 ===== ===== test accuracy is  0.932 ==========
========== 180 ===== ===== test accuracy is  0.932 ==========
========== 200 ===== ===== test accuracy is  0.93 ==========
========== 220 ===== ===== test accuracy is  0.94 ==========
========== 240 ===== ===== test accuracy is  0.918 ==========
========== 260 ===== ===== test accuracy is  0.934 ==========
========== 280 ===== ===== test accuracy is  0.93 ==========
========== 300 ===== ===== test accuracy is  0.934 ==========
========== 320 ===== ===== test accuracy is  0.934 ==========
========== 340 ===== ===== test accuracy is  0.93 ==========
========== 360 ===== ===== test accuracy is  0.944 ==========
========== 380 ===== ===== test accuracy is  0.938 ==========
========== 400 ===== ===== test accuracy is  0.92 ==========
========== 420 ===== ===== test accuracy is  0.936 ==========
========== 440 ===== ===== test accuracy is  0.948 ==========
========== 460 ===== ===== test accuracy is  0.934 ==========
========== 480 ===== ===== test accuracy is  0.938 ==========
========== 500 ===== ===== test accuracy is  0.916 ==========
========== 520 ===== ===== test accuracy is  0.916 ==========
========== 540 ===== ===== test accuracy is  0.928 ==========
========== 560 ===== ===== test accuracy is  0.936 ==========
========== 580 ===== ===== test accuracy is  0.942 ==========
========== 600 ===== ===== test accuracy is  0.922 ==========
========== 620 ===== ===== test accuracy is  0.94 ==========
========== 640 ===== ===== test accuracy is  0.94 ==========
========== 660 ===== ===== test accuracy is  0.96 ==========
========== 680 ===== ===== test accuracy is  0.938 ==========
========== 700 ===== ===== test accuracy is  0.936 ==========
========== 720 ===== ===== test accuracy is  0.94 ==========
========== 740 ===== ===== test accuracy is  0.946 ==========
========== 760 ===== ===== test accuracy is  0.946 ==========
========== 780 ===== ===== test accuracy is  0.948 ==========
========== 800 ===== ===== test accuracy is  0.95 ==========
========== 820 ===== ===== test accuracy is  0.948 ==========
========== 840 ===== ===== test accuracy is  0.95 ==========
========== 860 ===== ===== test accuracy is  0.94 ==========
========== 880 ===== ===== test accuracy is  0.956 ==========
========== 900 ===== ===== test accuracy is  0.944 ==========
========== 920 ===== ===== test accuracy is  0.948 ==========
========== 940 ===== ===== test accuracy is  0.95 ==========
========== 960 ===== ===== test accuracy is  0.944 ==========
========== 980 ===== ===== test accuracy is  0.94 ==========
========== 1000 ===== ===== test accuracy is  0.946 ==========
========== 1020 ===== ===== test accuracy is  0.952 ==========
========== 1040 ===== ===== test accuracy is  0.952 ==========
========== 1060 ===== ===== test accuracy is  0.944 ==========
========== 1080 ===== ===== test accuracy is  0.956 ==========
========== 1100 ===== ===== test accuracy is  0.96 ==========
========== 1120 ===== ===== test accuracy is  0.948 ==========
========== 1140 ===== ===== test accuracy is  0.942 ==========
========== 1160 ===== ===== test accuracy is  0.948 ==========
========== 1180 ===== ===== test accuracy is  0.944 ==========
========== 20 ===== ===== test accuracy is  0.952 ==========
========== 40 ===== ===== test accuracy is  0.96 ==========
========== 60 ===== ===== test accuracy is  0.948 ==========
========== 80 ===== ===== test accuracy is  0.954 ==========
========== 100 ===== ===== test accuracy is  0.948 ==========
========== 120 ===== ===== test accuracy is  0.948 ==========
========== 140 ===== ===== test accuracy is  0.958 ==========
========== 160 ===== ===== test accuracy is  0.942 ==========
========== 180 ===== ===== test accuracy is  0.948 ==========
========== 200 ===== ===== test accuracy is  0.952 ==========
========== 220 ===== ===== test accuracy is  0.952 ==========
========== 240 ===== ===== test accuracy is  0.95 ==========
========== 260 ===== ===== test accuracy is  0.966 ==========
========== 280 ===== ===== test accuracy is  0.96 ==========
========== 300 ===== ===== test accuracy is  0.956 ==========
========== 320 ===== ===== test accuracy is  0.96 ==========
========== 340 ===== ===== test accuracy is  0.956 ==========
========== 360 ===== ===== test accuracy is  0.956 ==========
========== 380 ===== ===== test accuracy is  0.954 ==========
========== 400 ===== ===== test accuracy is  0.96 ==========
========== 420 ===== ===== test accuracy is  0.966 ==========
========== 440 ===== ===== test accuracy is  0.96 ==========
========== 460 ===== ===== test accuracy is  0.954 ==========
========== 480 ===== ===== test accuracy is  0.968 ==========
========== 500 ===== ===== test accuracy is  0.958 ==========
========== 520 ===== ===== test accuracy is  0.958 ==========
========== 540 ===== ===== test accuracy is  0.962 ==========
========== 560 ===== ===== test accuracy is  0.968 ==========
========== 580 ===== ===== test accuracy is  0.958 ==========
========== 600 ===== ===== test accuracy is  0.952 ==========
========== 620 ===== ===== test accuracy is  0.95 ==========
========== 640 ===== ===== test accuracy is  0.964 ==========
========== 660 ===== ===== test accuracy is  0.962 ==========
========== 680 ===== ===== test accuracy is  0.96 ==========
========== 700 ===== ===== test accuracy is  0.962 ==========
========== 720 ===== ===== test accuracy is  0.964 ==========
========== 740 ===== ===== test accuracy is  0.958 ==========
========== 760 ===== ===== test accuracy is  0.96 ==========
========== 780 ===== ===== test accuracy is  0.972 ==========
========== 800 ===== ===== test accuracy is  0.962 ==========
========== 820 ===== ===== test accuracy is  0.968 ==========
========== 840 ===== ===== test accuracy is  0.964 ==========
========== 860 ===== ===== test accuracy is  0.96 ==========
========== 880 ===== ===== test accuracy is  0.964 ==========
========== 900 ===== ===== test accuracy is  0.96 ==========
========== 920 ===== ===== test accuracy is  0.96 ==========
========== 940 ===== ===== test accuracy is  0.97 ==========
========== 960 ===== ===== test accuracy is  0.956 ==========
========== 980 ===== ===== test accuracy is  0.966 ==========
========== 1000 ===== ===== test accuracy is  0.964 ==========
========== 1020 ===== ===== test accuracy is  0.964 ==========
========== 1040 ===== ===== test accuracy is  0.97 ==========
========== 1060 ===== ===== test accuracy is  0.974 ==========
========== 1080 ===== ===== test accuracy is  0.962 ==========
========== 1100 ===== ===== test accuracy is  0.97 ==========
========== 1120 ===== ===== test accuracy is  0.974 ==========
========== 1140 ===== ===== test accuracy is  0.978 ==========
========== 1160 ===== ===== test accuracy is  0.976 ==========
========== 1180 ===== ===== test accuracy is  0.974 ==========

在这里插入图片描述

本文来自互联网用户投稿，该文观点仅代表作者本人，不代表本站立场。本站仅提供信息存储空间服务，不拥有所有权，不承担相关法律责任。如若转载，请注明出处：http://www.mzph.cn/diannao/20322.shtml

如若内容造成侵权/违法违规/事实不符，请联系多彩编程网进行投诉反馈email:809451989@qq.com，一经查实，立即删除！

相关文章

商用未来何时来？软银揭示量子计算商业应用现状

商用未来何时来？软银揭示量子计算商业应用现状

内容来源：量子前哨（ID：Qforepost） 文丨沛贤/浪味仙 排版丨沛贤 深度好文：3000字丨10分钟阅读 摘要：软银（SoftBank）先进技术研究所正在积极推进量子计算商业应用，借助与…
阅读更多...
SpringCloud Feign用法

SpringCloud Feign用法

1.在目标应用的启动类上添加开启远程fein调用注解： 2.添加一个feign调用的interface FeignClient("gulimall-coupon") public interface CouponFeignService {PostMapping("/coupon/spubounds/save")R save(RequestBody SpuBondTo spuBounds);…
阅读更多...
随记-点选验证码（二）

随记-点选验证码（二）

之前写过一篇文章 随记-点选验证码 ，当时借助了 ddddocr 完成了ocr 识别，这篇文章算是对之前的补充。 本次更换了新的方案： 通过 ultralytics（YOLO8）训练自己的模型 吐槽一句：标注真是一件耗时的事情啊&am…
阅读更多...
鸿蒙实现汉字转拼音

鸿蒙实现汉字转拼音

1.使用三方库 pinyin-pro 地址：OpenHarmony三方库中心仓 亲测可用，一共三个关于 转pinyin的库，一个无法使用，另一个时间太久。 ohpm i pinyin-proimport { pinyin } from pinyin-pro;// 获取带音调拼音 pinyin(汉语拼音); // …
阅读更多...
【Java数据结构】详解LinkedList与链表（一）

【Java数据结构】详解LinkedList与链表（一）

🔒文章目录： 1.❤️❤️前言~🥳🎉🎉🎉 2.ArrayList的缺陷 3.链表的概念及结构 4.无头单向非循环链表的实现 4.1成员属性 4.2成员方法 createList display——打印链表 addFirst——头插 addLast…
阅读更多...
少样本学习与零样本学习：理解与应用

少样本学习与零样本学习：理解与应用

少样本学习与零样本学习：理解与应用 在现代机器学习领域中，少样本学习（Few-Shot Learning）和零样本学习（Zero-Shot Learning）正变得越来越重要。这些技术能够在数据稀缺的情况下有效地进行学习和推理&…
阅读更多...
2024就业寒潮下的挑战与机遇：能否守住饭碗，人工智能能否成为新春天？

2024就业寒潮下的挑战与机遇：能否守住饭碗，人工智能能否成为新春天？

前言 随着时代的飞速发展，2024年的就业市场迎来了前所未有的挑战。数以百万计的高校毕业生涌入市场，使得就业竞争愈发激烈。然而，在这股就业寒潮中，我们也看到了新的曙光——人工智能的崛起。这一新兴行业以其独特的魅力和巨大的…
阅读更多...
深入分析 Android Service (五)

深入分析 Android Service (五)

文章目录 深入分析 Android Service (五)1. 深入分析 Service 与 Activity 之间的通信2. Messenger 的内部工作原理2.1 服务端实现2.2 客户端实现 3. AIDL 的内部工作原理3.1 定义 AIDL 接口3.2 服务端实现3.3 客户端实现 4. Service 的优化建议和最佳实践4.1 异步操作4.2 资源…
阅读更多...
【Linux】权限的概念

【Linux】权限的概念

1.Linux权限的概念 Linux下有两种用户：超级用户（root）、普通用户。 超级用户：可以再linux系统下做任何事情，不受权限限制 普通用户：在linux下做有限的事情，受权限设置。 windows下也有超级用户…
阅读更多...
Object.entries方法的使用

Object.entries方法的使用

Object.entries() 方法返回一个给定对象自身可枚举属性的键值对数组。 有以下需求&#xff1a; let cpuData reactive([{ label: 总量, content: test },{ label: 已使用, content: test },{ label: 未使用, content: test } ])<el-form label-position"left" l…
阅读更多...
环卫车北斗GPS视频监控定位解决方案的应用与优势

环卫车北斗GPS视频监控定位解决方案的应用与优势

一、引言 随着城市化进程的加快&#xff0c;环卫车作为城市环境卫生的重要保障力量&#xff0c;其运行效率与安全性直接关系到城市形象与居民生活品质。然而&#xff0c;传统的环卫车管理模式往往存在信息不对称、调度不合理、行驶不规范等问题&#xff0c;导致城市道路污染和…
阅读更多...
微信小程序对接发货功能

微信小程序对接发货功能

注&#xff1a;微信小程序对接发货功能 文档地址&#xff1a;https://developers.weixin.qq.com/miniprogram/dev/platform-capabilities/business-capabilities/order-shipping/order-shipping.html php代码 common.php use think\Config; use think\Db; use fast\Http; us…
阅读更多...
LabVIEW远程开发与调试

LabVIEW远程开发与调试

在现代项目开发中&#xff0c;远程开发与调试已经成为一种常见的模式&#xff0c;特别是在使用LabVIEW进行工程项目时。本文将详细分析LabVIEW远程开发与调试的优缺点&#xff0c;并从多个角度说明如何建议客户采用这种方式&#xff0c;以提高项目效率和质量。 优点 灵活性和便…
阅读更多...
Linux【安全 02】OpenSSH漏洞修复（离线升级最新版本流程）网盘分享3个安装包+26个离线依赖

Linux【安全 02】OpenSSH漏洞修复（离线升级最新版本流程）网盘分享3个安装包+26个离线依赖

OpenSSH离线升级最新版本流程 1. 漏洞信息2. 环境说明3.依赖安装3.1 在线安装3.2 离线安装 4.备份卸载4.1 备份4.2 卸载旧版本 5.安装5.1 zlib5.2 ssl5.3 openssh5.3.1 安装5.3.2 配置 6.脚本整理7.文件资源 本文仅针对CentOS7.8版本&#xff0c;其他版本未测试&#xff0c;安装…
阅读更多...
GSM信令流程(附着、去附着、PDP激活、修改流程)

GSM信令流程(附着、去附着、PDP激活、修改流程)

1、联合附着流程 附着包括身份认证、鉴权等 2、去附着流程 用户发起去附着 SGSN发起去附着 HLR发起去附着 GSSN使用S4发起去附着 3、Activation Procedures(PDP激活流程) 4、PDP更新或修改流程 5、Deactivate PDP Context 6、RAU(Routeing Area Update)流程 7、鉴权加…
阅读更多...
生成式AI，在云端的绽放与盛开

生成式AI，在云端的绽放与盛开

编辑&#xff1a;阿冒 设计&#xff1a;沐由 毫无疑问&#xff0c;生成式AI已然成为当今技术发展和应用创新的重要引擎之一。 过去的一年多时间里&#xff0c;我们每个人都在目睹和见证着生成式AI是如何以移山倒海的力量&#xff0c;为诸多行业带来革命性乃至颠覆性的变革&…
阅读更多...
新版校园跑腿外卖独立版+APP+小程序前端外卖配送平台源码

新版校园跑腿外卖独立版+APP+小程序前端外卖配送平台源码

同城校园跑腿外卖配送平台源码&#xff0c;这套目前全网还没有人分享过&#xff0c;这个是开源的&#xff0c;所以没有任何问题了&#xff0c;这套源码非常吊&#xff0c;支持自定义diy 你可以设计你的页面&#xff0c;设计你自己的风格&#xff0c;支持多校园&#xff0c;独立…
阅读更多...
深度学习中的模型架构详解：RNN、LSTM、TextCNN和Transformer

深度学习中的模型架构详解：RNN、LSTM、TextCNN和Transformer

深度学习中的模型架构详解&#xff1a;RNN、LSTM、TextCNN和Transformer 文章目录 深度学习中的模型架构详解&#xff1a;RNN、LSTM、TextCNN和Transformer循环神经网络 (RNN)RNN的优点RNN的缺点RNN的代码实现 长短期记忆网络 (LSTM)LSTM的优点LSTM的缺点LSTM的代码实现 TextCN…
阅读更多...
mac电脑安卓设备文件传输助手：MacDroid pro 中文激活版

mac电脑安卓设备文件传输助手：MacDroid pro 中文激活版

MacDroid Pro是一款专为Mac电脑和Android设备设计的软件&#xff0c;旨在简化两者之间的文件传输和数据管理&#xff0c;双向文件传输&#xff1a;支持从Mac电脑向Android设备传输文件&#xff0c;也可以将Android设备上的文件轻松传输到Mac电脑上。完整的文件访问和管理&#…
阅读更多...
机器学习笔记 - PyTorch 分布式训练概览

机器学习笔记 - PyTorch 分布式训练概览

一、简述 对于大规模的数据集，只能进行分布式训练，分布式训练会尽可能的利用我们的算力，使模型训练更加高效。PyTorch提供了Data Parallel包，它可以实现单机、多GPU并行。 PyTorch 数据并行模块的内部工作原理 上面的图像说明了PyTorch 如何在单个系统中利用多个 G…
阅读更多...
推荐文章
最新文章

Copyright @ 2022~2023