在课堂中学习了Resnet模型，并根据老师所给程序在ModelArts创建Notebook进行运行。以下是一些学习实践过程及结果的记录与分享：1.Resent的实现（1）导入工具包（2）定义Resent模型（3）获得数据集（4）定义训练过程（基于训练集）（5）定义测试验证（基于验证集）（6）定义超参数（7）开始训练2.运行及测试结果（1）根据所给程序，初始设置batch_size = 1024；learning_rate=0.0001。此时，train loss 和 test loss 随 Epoch 的部分变化结果如下所示：（最终准确率在30％左右）Files already downloaded and verifiedFiles already downloaded and verified================ epoch: 0 ================Epoch: 0, Step 0, Loss: 4.738948345184326Epoch: 0, Step 10, Loss: 4.255447864532471Epoch: 0, Step 20, Loss: 3.9827303886413574Epoch: 0, Step 30, Loss: 3.7523534297943115Epoch: 0, Step 40, Loss: 3.5549187660217285Test Accuracy: 16.689999999999998%================ epoch: 1 ================Epoch: 1, Step 0, Loss: 3.417269706726074Epoch: 1, Step 10, Loss: 3.238386392593384Epoch: 1, Step 20, Loss: 3.1039466857910156Epoch: 1, Step 30, Loss: 2.972597599029541Epoch: 1, Step 40, Loss: 2.856231451034546Test Accuracy: 26.08%================ epoch: 2 ================Epoch: 2, Step 0, Loss: 2.8012454509735107Epoch: 2, Step 10, Loss: 2.572690725326538Epoch: 2, Step 20, Loss: 2.436326503753662Epoch: 2, Step 30, Loss: 2.320054531097412Epoch: 2, Step 40, Loss: 2.2200241088867188Test Accuracy: 29.599999999999998%================ epoch: 3 ================Epoch: 3, Step 0, Loss: 2.1936466693878174Epoch: 3, Step 10, Loss: 1.915575385093689Epoch: 3, Step 20, Loss: 1.7562488317489624Epoch: 3, Step 30, Loss: 1.6630589962005615Epoch: 3, Step 40, Loss: 1.5405477285385132Test Accuracy: 28.74%================ epoch: 4 ================Epoch: 4, Step 0, Loss: 1.4985536336898804Epoch: 4, Step 10, Loss: 1.2820249795913696Epoch: 4, Step 20, Loss: 1.1182427406311035Epoch: 4, Step 30, Loss: 1.054147481918335Epoch: 4, Step 40, Loss: 1.022835612297058Test Accuracy: 26.1%================ epoch: 5 ================Epoch: 5, Step 0, Loss: 0.9999890923500061Epoch: 5, Step 10, Loss: 0.885351300239563Epoch: 5, Step 20, Loss: 0.794113039970398Epoch: 5, Step 30, Loss: 0.7815513014793396Epoch: 5, Step 40, Loss: 0.7719193696975708Test Accuracy: 23.71%================ epoch: 6 ================Epoch: 6, Step 0, Loss: 0.807374894618988Epoch: 6, Step 10, Loss: 0.7296508550643921Epoch: 6, Step 20, Loss: 0.6557181477546692Epoch: 6, Step 30, Loss: 0.5987974405288696Epoch: 6, Step 40, Loss: 0.5647777318954468Test Accuracy: 24.98%================ epoch: 7 ================Epoch: 7, Step 0, Loss: 0.5228223204612732Epoch: 7, Step 10, Loss: 0.521501362323761Epoch: 7, Step 20, Loss: 0.4148188531398773Epoch: 7, Step 30, Loss: 0.3891873359680176Epoch: 7, Step 40, Loss: 0.3635284900665283Test Accuracy: 25.740000000000002%================ epoch: 8 ================Epoch: 8, Step 0, Loss: 0.35147446393966675Epoch: 8, Step 10, Loss: 0.27659207582473755Epoch: 8, Step 20, Loss: 0.2709190845489502Epoch: 8, Step 30, Loss: 0.2323095202445984Epoch: 8, Step 40, Loss: 0.23784960806369781Test Accuracy: 26.39%================ epoch: 9 ================Epoch: 9, Step 0, Loss: 0.22373342514038086Epoch: 9, Step 10, Loss: 0.2217608392238617Epoch: 9, Step 20, Loss: 0.15514764189720154Epoch: 9, Step 30, Loss: 0.16614097356796265Epoch: 9, Step 40, Loss: 0.1333569884300232Test Accuracy: 27.24%================ epoch: 10 ================Epoch: 10, Step 0, Loss: 0.1310970038175583Epoch: 10, Step 10, Loss: 0.10795009136199951Epoch: 10, Step 20, Loss: 0.09246394783258438Epoch: 10, Step 30, Loss: 0.10038184374570847Epoch: 10, Step 40, Loss: 0.09417460858821869Test Accuracy: 27.639999999999997%================ epoch: 11 ================Epoch: 11, Step 0, Loss: 0.07077963650226593Epoch: 11, Step 10, Loss: 0.07666710764169693Epoch: 11, Step 20, Loss: 0.0636228695511818Epoch: 11, Step 30, Loss: 0.04714050889015198Epoch: 11, Step 40, Loss: 0.06018964573740959Test Accuracy: 28.71%================ epoch: 12 ================Epoch: 12, Step 0, Loss: 0.05108068883419037Epoch: 12, Step 10, Loss: 0.04214892536401749Epoch: 12, Step 20, Loss: 0.03265600651502609Epoch: 12, Step 30, Loss: 0.03236588463187218Epoch: 12, Step 40, Loss: 0.03212810680270195Test Accuracy: 28.99%================ epoch: 13 ================Epoch: 13, Step 0, Loss: 0.02681496925652027Epoch: 13, Step 10, Loss: 0.022608093917369843Epoch: 13, Step 20, Loss: 0.01608162187039852Epoch: 13, Step 30, Loss: 0.03622477874159813Epoch: 13, Step 40, Loss: 0.026449179276823997Test Accuracy: 29.470000000000002%================ epoch: 14 ================Epoch: 14, Step 0, Loss: 0.01653885468840599Epoch: 14, Step 10, Loss: 0.010875535197556019Epoch: 14, Step 20, Loss: 0.011256149038672447Epoch: 14, Step 30, Loss: 0.029273658990859985Epoch: 14, Step 40, Loss: 0.032421231269836426Test Accuracy: 29.049999999999997%================ epoch: 15 ================Epoch: 15, Step 0, Loss: 0.013880199752748013Epoch: 15, Step 10, Loss: 0.010487137362360954Epoch: 15, Step 20, Loss: 0.01054014265537262Epoch: 15, Step 30, Loss: 0.02424754574894905Epoch: 15, Step 40, Loss: 0.012607093900442123Test Accuracy: 29.62%================ epoch: 16 ================Epoch: 16, Step 0, Loss: 0.008765900507569313Epoch: 16, Step 10, Loss: 0.007048243656754494Epoch: 16, Step 20, Loss: 0.008071326650679111Epoch: 16, Step 30, Loss: 0.016480540856719017Epoch: 16, Step 40, Loss: 0.011516512371599674Test Accuracy: 30.0%================ epoch: 17 ================Epoch: 17, Step 0, Loss: 0.0064095971174538136Epoch: 17, Step 10, Loss: 0.004685310181230307Epoch: 17, Step 20, Loss: 0.004435277543962002Epoch: 17, Step 30, Loss: 0.017189662903547287Epoch: 17, Step 40, Loss: 0.011844292283058167Test Accuracy: 30.04%================ epoch: 18 ================Epoch: 18, Step 0, Loss: 0.004041185602545738Epoch: 18, Step 10, Loss: 0.004346000961959362Epoch: 18, Step 20, Loss: 0.008933650329709053Epoch: 18, Step 30, Loss: 0.012093418277800083Epoch: 18, Step 40, Loss: 0.008453783579170704Test Accuracy: 30.020000000000003%================ epoch: 19 ================Epoch: 19, Step 0, Loss: 0.003529253415763378Epoch: 19, Step 10, Loss: 0.002985781291499734Epoch: 19, Step 20, Loss: 0.0027229641564190388Epoch: 19, Step 30, Loss: 0.021751798689365387Epoch: 19, Step 40, Loss: 0.00738913007080555Test Accuracy: 30.099999999999998%================ epoch: 20 ================Epoch: 20, Step 0, Loss: 0.003075427608564496Epoch: 20, Step 10, Loss: 0.0026734406128525734Epoch: 20, Step 20, Loss: 0.00244825123809278Epoch: 20, Step 30, Loss: 0.007915157824754715Epoch: 20, Step 40, Loss: 0.006863289512693882Test Accuracy: 30.3%================ epoch: 21 ================Epoch: 21, Step 0, Loss: 0.002420809818431735Epoch: 21, Step 10, Loss: 0.002310082782059908Epoch: 21, Step 20, Loss: 0.002210301114246249Epoch: 21, Step 30, Loss: 0.018848098814487457Epoch: 21, Step 40, Loss: 0.006330176256597042Test Accuracy: 30.240000000000002%================ epoch: 22 ================Epoch: 22, Step 0, Loss: 0.002132808556780219Epoch: 22, Step 10, Loss: 0.002030259696766734Epoch: 22, Step 20, Loss: 0.0019014010904356837Epoch: 22, Step 30, Loss: 0.005129395984113216Epoch: 22, Step 40, Loss: 0.005537337623536587Test Accuracy: 30.270000000000003%================ epoch: 23 ================Epoch: 23, Step 0, Loss: 0.0019582901149988174Epoch: 23, Step 10, Loss: 0.0018943308386951685Epoch: 23, Step 20, Loss: 0.0017575022066012025Epoch: 23, Step 30, Loss: 0.016244899481534958Epoch: 23, Step 40, Loss: 0.005317316856235266Test Accuracy: 30.34%================ epoch: 24 ================Epoch: 24, Step 0, Loss: 0.001790825743228197Epoch: 24, Step 10, Loss: 0.0017201013397425413Epoch: 24, Step 20, Loss: 0.0016320045106112957Epoch: 24, Step 30, Loss: 0.004616044461727142Epoch: 24, Step 40, Loss: 0.005004520528018475Test Accuracy: 30.37%================ epoch: 25 ================Epoch: 25, Step 0, Loss: 0.001665924210101366Epoch: 25, Step 10, Loss: 0.001624491298571229Epoch: 25, Step 20, Loss: 0.0015230735298246145Epoch: 25, Step 30, Loss: 0.012112809345126152Epoch: 25, Step 40, Loss: 0.004875780548900366Test Accuracy: 30.349999999999998%================ epoch: 26 ================Epoch: 26, Step 0, Loss: 0.001563299330882728Epoch: 26, Step 10, Loss: 0.0015107914805412292Epoch: 26, Step 20, Loss: 0.0014351849677041173Epoch: 26, Step 30, Loss: 0.004178904928267002Epoch: 26, Step 40, Loss: 0.0048788730055093765Test Accuracy: 30.36%================ epoch: 27 ================Epoch: 27, Step 0, Loss: 0.0014841527445241809Epoch: 27, Step 10, Loss: 0.0014560072449967265Epoch: 27, Step 20, Loss: 0.0013575487537309527Epoch: 27, Step 30, Loss: 0.012932670302689075Epoch: 27, Step 40, Loss: 0.0047125499695539474Test Accuracy: 30.31%================ epoch: 28 ================Epoch: 28, Step 0, Loss: 0.0013955835020169616Epoch: 28, Step 10, Loss: 0.0013590363087132573Epoch: 28, Step 20, Loss: 0.0012899343855679035Epoch: 28, Step 30, Loss: 0.003891129745170474Epoch: 28, Step 40, Loss: 0.004475367721170187Test Accuracy: 30.3%================ epoch: 29 ================Epoch: 29, Step 0, Loss: 0.0013218162348493934Epoch: 29, Step 10, Loss: 0.0012971273390576243Epoch: 29, Step 20, Loss: 0.0012194958981126547Epoch: 29, Step 30, Loss: 0.010904560796916485Epoch: 29, Step 40, Loss: 0.004483921453356743Test Accuracy: 30.349999999999998%================ epoch: 30 ================Epoch: 30, Step 0, Loss: 0.0012635505991056561Epoch: 30, Step 10, Loss: 0.0012238533236086369Epoch: 30, Step 20, Loss: 0.001171094598248601Epoch: 30, Step 30, Loss: 0.003731002099812031Epoch: 30, Step 40, Loss: 0.004489581100642681Test Accuracy: 30.44%================ epoch: 31 ================Epoch: 31, Step 0, Loss: 0.0012064295588061213Epoch: 31, Step 10, Loss: 0.0011814942117780447Epoch: 31, Step 20, Loss: 0.0011084245052188635Epoch: 31, Step 30, Loss: 0.010918247513473034Epoch: 31, Step 40, Loss: 0.004249677062034607Test Accuracy: 30.44%================ epoch: 32 ================Epoch: 32, Step 0, Loss: 0.0011464636772871017Epoch: 32, Step 10, Loss: 0.0011157296830788255Epoch: 32, Step 20, Loss: 0.0010655020596459508Epoch: 32, Step 30, Loss: 0.0035718355793505907Epoch: 32, Step 40, Loss: 0.004267847631126642Test Accuracy: 30.509999999999998%================ epoch: 33 ================Epoch: 33, Step 0, Loss: 0.0010997997596859932Epoch: 33, Step 10, Loss: 0.0010838627349585295Epoch: 33, Step 20, Loss: 0.0010143746621906757Epoch: 33, Step 30, Loss: 0.010117175057530403Epoch: 33, Step 40, Loss: 0.004162825644016266Test Accuracy: 30.43%================ epoch: 34 ================Epoch: 34, Step 0, Loss: 0.0010507091647014022Epoch: 34, Step 10, Loss: 0.0010267348261550069Epoch: 34, Step 20, Loss: 0.0009758920641615987Epoch: 34, Step 30, Loss: 0.003296484239399433Epoch: 34, Step 40, Loss: 0.004471142776310444Test Accuracy: 30.580000000000002%================ epoch: 35 ================Epoch: 35, Step 0, Loss: 0.0010124717373400927Epoch: 35, Step 10, Loss: 0.0009983875788748264Epoch: 35, Step 20, Loss: 0.0009379943949170411Epoch: 35, Step 30, Loss: 0.010011741891503334Epoch: 35, Step 40, Loss: 0.004122457932680845Test Accuracy: 30.45%================ epoch: 36 ================Epoch: 36, Step 0, Loss: 0.0009714484331198037Epoch: 36, Step 10, Loss: 0.0009493435500189662Epoch: 36, Step 20, Loss: 0.0009067429346032441Epoch: 36, Step 30, Loss: 0.003059440990909934Epoch: 36, Step 40, Loss: 0.003779624355956912Test Accuracy: 30.59%================ epoch: 37 ================Epoch: 37, Step 0, Loss: 0.0009371026535518467Epoch: 37, Step 10, Loss: 0.0009284141124226153Epoch: 37, Step 20, Loss: 0.0008677949081175029Epoch: 37, Step 30, Loss: 0.009749548509716988Epoch: 37, Step 40, Loss: 0.004097274504601955Test Accuracy: 30.44%================ epoch: 38 ================Epoch: 38, Step 0, Loss: 0.0008963147993199527Epoch: 38, Step 10, Loss: 0.0008796583279035985Epoch: 38, Step 20, Loss: 0.0008349142735823989Epoch: 38, Step 30, Loss: 0.0028689114842563868Epoch: 38, Step 40, Loss: 0.0037223438266664743Test Accuracy: 30.53%================ epoch: 39 ================Epoch: 39, Step 0, Loss: 0.0008696657023392618Epoch: 39, Step 10, Loss: 0.0008523689466528594Epoch: 39, Step 20, Loss: 0.0008050688775256276Epoch: 39, Step 30, Loss: 0.009185832925140858Epoch: 39, Step 40, Loss: 0.0039543770253658295Test Accuracy: 30.54%================ epoch: 40 ================Epoch: 40, Step 0, Loss: 0.0008351270807906985Epoch: 40, Step 10, Loss: 0.0008215241250582039Epoch: 40, Step 20, Loss: 0.0007779219304211438Epoch: 40, Step 30, Loss: 0.0026640305295586586Epoch: 40, Step 40, Loss: 0.003869910491630435Test Accuracy: 30.620000000000005%（2）尝试提高最终准确率。选择通过改变一次训练的样本数目以及学习率，以提高最终接近稳定状态下的准确率。其他可以提高准确率的方法有：使用dropout、batch_nomalization等。经过多次尝试，选取以下两种情况进行记录：（最终将接近稳定状态下的准确率从30％提高至50％左右）（2-1）修改一次batch_size = 2048；learning_rate = 0.001此时，train loss 和 test loss 随 Epoch 的部分变化结果如下所示：（最终准确率在45％左右）运行发现在60次训练左右准确率趋于稳定：（截取最后几次训练数据进行展示）================ epoch: 56 ================Epoch: 56, Step 0, Loss: 7.18005103408359e-05Epoch: 56, Step 10, Loss: 7.413411367451772e-05Epoch: 56, Step 20, Loss: 0.001145041431300342Test Accuracy: 45.03%================ epoch: 57 ================Epoch: 57, Step 0, Loss: 6.976973236305639e-05Epoch: 57, Step 10, Loss: 7.286847539944574e-05Epoch: 57, Step 20, Loss: 0.0006657794583588839Test Accuracy: 45.050000000000004%================ epoch: 58 ================Epoch: 58, Step 0, Loss: 6.862186273792759e-05Epoch: 58, Step 10, Loss: 7.077038753777742e-05Epoch: 58, Step 20, Loss: 0.0011522056302055717Test Accuracy: 45.0%================ epoch: 59 ================Epoch: 59, Step 0, Loss: 6.702987593598664e-05Epoch: 59, Step 10, Loss: 6.995741568971425e-05Epoch: 59, Step 20, Loss: 0.0006697141216136515Test Accuracy: 45.129999999999995%================ epoch: 60 ================Epoch: 60, Step 0, Loss: 6.572520942427218e-05Epoch: 60, Step 10, Loss: 6.784420111216605e-05Epoch: 60, Step 20, Loss: 0.0010953237069770694Test Accuracy: 45.08%（2-2）修改一次batch_size = 4096；learning_rate = 0.002此时，train loss 和 test loss 随 Epoch 的部分变化结果如下所示：（最终准确率在49％左右）运行发现在65次训练左右准确率趋于稳定：（截取最后几次训练数据进行展示）================ epoch: 61 ================Epoch: 61, Step 0, Loss: 0.00024627658422105014Epoch: 61, Step 10, Loss: 0.0005478289094753563Test Accuracy: 49.57%================ epoch: 62 ================Epoch: 62, Step 0, Loss: 0.00023818077170290053Epoch: 62, Step 10, Loss: 0.0005578603013418615Test Accuracy: 49.559999999999995%================ epoch: 63 ================Epoch: 63, Step 0, Loss: 0.00023028196301311255Epoch: 63, Step 10, Loss: 0.0005240143509581685Test Accuracy: 49.559999999999995%================ epoch: 64 ================Epoch: 64, Step 0, Loss: 0.0002232123661087826Epoch: 64, Step 10, Loss: 0.0005316751776263118Test Accuracy: 49.54%================ epoch: 65 ================Epoch: 65, Step 0, Loss: 0.00021623082284349948Epoch: 65, Step 10, Loss: 0.0005059240502305329Test Accuracy: 49.64%

一、简介本项目主要运用华为云 EI 的 ModelArts 的自动学习以及云对象存储的 OBS，实现简单的垃圾分类系统。二、内容描述本垃圾分类图像识别系统主要通过创建图像分类自动学习项目，进行数据标注，进行自动训练和部署测试，再到最后的结束测试。 三、主要流程四、图像分类任务介绍ModelArts 服务之自动学习图像分类项目，是对图像进行检测分类。添加图片并对图像进行分类标注，每个分类识别一种类型的图像。完成图片标注后开始自动训练，即可快速生成图像分类模型。可应用于商品的自动识别、运输车辆种类识别和残次品的自动检测。例如质量检查的场景，则可以上传产品图片，将图片标注“合格”、“不合格”，通过训练部署模型，实现产品的质检。五、系统创建1、创建项目2、添加图片3、数据标注进行“一次性快餐盒-其他垃圾” 的数据标注。 先将左下角的数字选择为 45， 点击图片选择同类的图片（一次可以选择一张或者多张），在标签名栏填写当前选择图片的标签（已有的标签可以直接选择） ， 输入标签名， 点击确定。4、自动训练5、部署测试预测结果：

3.0版本在部分W11系统上安装并激活后之出现这个提示，还有OCR的ID和Key运行提示接口有问题，ID和Key没有输错

重庆工程学院HCSD活动复盘总结HCSD名称：重庆工程学院校园大使姓名：肖宏城所在学校：重庆工程学院活动背景介绍活动目的及意义为进一步加强计算机专业学生的综合实力，增强计算机专业的技术水平，提高社会竞争力，以学致用，学用结合，重庆工程学院联合多专业社团开展了“华为云HCSD码上服务”活动。邀请华为云专家到场解读行业最新动态，向同学们分享前沿技术知识，热门的实验分享，场景化的动手实践，让大家充分体验到华为云产品的技术魅力。通过宣讲，以及教师团队的协助，给未接触和接触较浅的同学和老师进行宣讲，让其了解和深入学习我们华为云的一些技术，了解我们华为云的产品以及好处，加大师生对华为云的认知度，认可度以及能通过此次活动吸取更多优秀的人才进入到我们的组织。活动预期目标活动宣传辐射人数全校计算机专业同学、活动参与人数力争200余人，通过介绍华为云的技术以及产品，吸引更多的开发者使用华为API、开发工具以及一些云服务，了解华为云带来的功能以及对于自身带来的好处。活动效果目标达成情况：宣传实际辐射人数达到300余人、活动实际参与人数120余人、体验平台人数预计80人活动现场情况：活动现场学生积极进行互动和提问，对于未来个人就业发展以及行业发展趋势有一定的自我认知，扩展了对于计算机行业的多职务的认识；现场氛围良好，工作人员细心的对学生进行引导，活跃现场气氛，积极组织现场秩序，带动现场的趣味性；华为云DTSE专家热情的回答大家的问题，给大家讲解华为云以及华为的业务范围，打破大众化认知。优缺点及改进方案活动亮点现场活动氛围良好，学生富有积极性，踊跃参加开发者认证申请，积极提出华为云存在的一些不足点，帮助华为云后续的发展。活动不足及改进方案对于宣传的方式以及途径有待改进，活动的整个申请到开展以及结束过程含有多方面沟通问题，对于活动开展的校方人员等时间安排和沟通上存在不足，校方的人员因为各种活动未到场进行一定的积极作用引导，对于一些因素仍未考虑到位。活动现场照片资源的投入及使用情况资源名称数量单价总价已发放数量结余数量荣耀体脂秤511055041数据线20011220096104定制T恤1550750150ModelArts人工智能应用开发指南30501500273帆布包1525375132