20行代码：Serverless架构下用Python可以轻松完成图片分类还是图片内容是另外一个API？首先推荐一个图像相关的库：ImageAI通过官方代码，我们可以看到一个简单的demo：fromimageai.PredictionimportImagePredictionimportosexecution_path=os.getcwd()prediction=ImagePrediction()prediction.setModelTypeAsResNet()prediction.setModelPath(os.path.join(execution_path,"resnet50_weights_tf_dim_ordering_tf_kernels.h5"))prediction.loadModel()预测,probabilities=prediction.predictImage(os.path.join(execution_path,"1.jpg"),result_count=5)foreachPrediction,eachProbabilityinzip(predictions,probabilities):print(eachPrediction+":"+eachProbability)通过这个Demo，我们可以考虑将这个模块部署到CloudFunction：首先我们在本地创建一个Python项目：mkdirimageDemo然后新建一个文件：vimindex.pyfromimageai.Prediction导入ImagePredictionimportos,base64,randomexecution_path=os.getcwd()prediction=ImagePrediction()prediction.setModelTypeAsSqueezeNet()prediction.setModelPath(os.path.join(execution_pat小时，“平方米uezenet_weights_tf_dim_ordering_tf_kernels.h5"))prediction.loadModel()defmain_handler(event,context):imgData=base64.b64decode(event["body"])fileName='/tmp/'+"".join(random.sample('zyxwvutsrqponmlkjihgfedcba',5))withopen(fileName,'wb')asf:f.write(imgData)resultData={}预测，概率=prediction.predictImage(fileName,result_count=5)foreachPrediction,eachProbabilityinzip(predictions,probabilities):resultData[eachPrediction]=eachProbabilityreturnresultData已创建，我们需要下载我们依赖的模型：-SqueezeNet（文件大小：4.82MB，预测时间最短，精度适中）-MicrosoftResearch的ResNet50（文件大小：98MB，更快的预测时间，更高的准确度）-GoogleBrain团队的InceptionV3（文件大小：91.6MB，预测时间更慢，准确度更高）-FacebookAIResearch的DenseNet121（文件大小：31.6MB，预测时间更慢，准确度最高）我们先用第一个SqueezeNet来测试一下：复制模型文件添加官方文档中的ress：直接使用wget安装：wgethttps://github.com/OlafenwaMoses/ImageAI/releases/download/1.0/squeezenet_weights_tf_dim_ordering_tf_kernels.h5接下来我们需要安装依赖是的，这里好像安装了很多内容：而且其中一些依赖需要编译，这就需要我们在centos+python2.7/3.6版本下打包，非常复杂，尤其是mac/windows用户，伤不起所以这个时候直接使用我之前的打包地址：直接下载解压，然后放到自己的项目中：最后一步，我们创建serverless.yamlimageDemo:component:"@serverless/tencent-scf"inputs:name:imageDemocodeUri:./handler:index.main_handlerruntime:Python3.6region:ap-guangzhoudescription:图像识别/分类DemomemorySize:256timeout:10events:-apiw:name:imageDemo_apigw_serviceparameters:protocols:-httpserviceName:serverlessdescription:Imagerecognition/classificationDemoAPI环境:releaseendpoints:-path:/imagemethod:ANY完成后，执行我们的sls--debug部署，部署过程中会有扫码登录，等待之后登录，完成后，我们可以复制生成的URL：通过Python语言测试，url就是我们刚刚复制的+/image：importurllib.requestimportbase64withopen("1.jpg",'rb')asf:base64_data=base64.b64encode(f.read())s=base64_data.decode()url='http://service-9p7hbgvg-1256773370.gz.apigw.tencentcs.com/release/image'print(urllib.request.urlopen(urllib.request.request(url=url,data=s.encode("utf-8"))).read().decode("utf-8"))通过网络搜索一张图片，比如我找到了这个：gettherunningresult：{"cheetah":83.12643766403198,"Irish_terrier":2.315458096563816,"lion":1.8476998433470726,"teddy":1.6655176877975464,"baboon":1.5562783926725388}将代码修改一下，进行一下简单的耗时测试：importurllib.requestimportbase64,timeforiinrange(0,10):start_time=time.time()withopen("1.jpg",'rb')asf:base64_data=base64.b64encode(f.read())s=base64_数据。decode()url='http://service-hh53d8yz-1256773370.bj.apigw.tencentcs.com/release/test'print(urllib.request.urlopen(urllib.request.Request(url=url,data=s.encode("utf-8")).read().decode("utf-8"))print("cost:",time.time()-start_time)output:{"猎豹":83.12643766403198,"Irish_terrier“：2.315458096563816，“狮子”：1.8476998433470726，“teddy”：1.66655176877975464，baboon，“baboon”：1.55562783926726726725388}rrier":2.315458096563816,"lion":1.8476998433470726,"teddy":1.6655176877975464,"baboon":1.5562783926725388}cost:1.1259253025054932{"cheetah":83.12643766403198,"Irish_terrier":2.315458096563816,"lion":1.8476998433470726,"teddy":1.6655176877975464,"baboon":1.5562783926725388}cost:1.3322770595550537{"cheetah":83.12643766403198,"Irish_terrier":2.315458096563816,"lion":1.8476998433470726,"teddy":1.6655176877975464,"baboon":1.5562783926725388}cost:1.3562259674072266{"cheetah":83.12643766403198,"Irish_terrier":2.315458096563816,"lion":1.8476998433470726,"teddy":1.6655176877975464,"baboon":1.5562783926725388}cost:1.0180821418762207{"cheetah":83.12643766403198,"Irish_terrier":2.315458096563816,"lion":1.8476998433470726,"teddy"：1.665517687775464，“狒狒”：1.5562783926725388}费用：1.429067134857777{“Cheetah”：83.1264376403198，“176877975464,"baboon":1.5562783926725388}cost:1.5917718410491943{"cheetah":83.12643766403198,"Irish_terrier":2.315458096563816,"lion":1.8476998433470726,"teddy":1.6655176877975464,"baboon":1.5562783926725388}cost:1.1727900505065918{"cheetah":83.12643766403198,"Irish_terrier":2.315458096563816,"lion":1.8476998433470726,"teddy":1.6655176877975464,"baboon":1.5562783926725388}cost:2.962592840194702{"cheetah":83.12643766403198,"Irish_terrier":2.315458096563816,"lion":1.8476998433470726,"teddy":1.6655176877975464,"baboon":1.5562783926725388}cost:1.2248001098632812这个数据，整体性能基本在我可以接受的范围内至此，我们就完成了通过Serverless架构搭建的Python版图像识别/分类工具。
