omega|ml-DataOps&MLOpsforhumas
withjustasiglelieofcodeyouca
deploymachielearigmodelsstraightfromJupyterNotebook(orayothercode)implemetdatapipeliesquickly,withoutmemorylimitatio,allfromaPadas-likeAPIservemodelsaddatafromaeasytouseRESTAPIFurther,omega|mlisthefastestwayto
scalemodeltraiigotheicludedscalablepure-Pythocomputecluster,oSparkorayothercloudcollaborateodatascieceprojectseasily,sharigJupyterNotebooksdeploybeautifuldashboardsrightfromyourJupyterNotebook,usigdashserveLiksDocumetatio:https://omegaml.github.io/omegaml/Cotributios:https://bit.ly/omegaml-cotributeGetstartedi<5miutesStarttheomega|mlserverrightfromyourlaptoporvirtualmachie
$wgethttps://raw.githubusercotet.com/omegaml/omegaml/master/docker-compose.yml$docker-composeup-dJupyterNotebookisimmediatelyavailableathttps://localhost:8899(omegamlisfutologi).Ayotebookyoucreatewillautomaticallybestoreditheitegratedomega|mldatabase,makigcollaboratioabreeze.TheRESTAPIisavailableathttps://localhost:5000.
AlreadyhaveaPythoeviromet(e.g.JupyterNotebook)?Leveragethepowerofomega|mlbyistalligasfollows:
#assumigyouhavestartedtheserverasperabove$pipistallomega|mlExamplesGetmoreiformatioathttps://omegaml.github.io/omegaml/
#trasparetlystorePadasSeriesadDataFramesorayPythoobjectom.datasets.put(df,'stats')om.datasets.get('stats',sales__gte=100)#trasparetlystoreadgetmodelsclf=LogisticRegressio()om.models.put(clf,'forecast')clf=om.models.get('forecast')#ruadscalemodelsdirectlyotheitegratedPythoorSparkcomputeclusterom.rutime.model('forecast').fit('stats[^sales]','stats[sales]')om.rutime.model('forecast').predict('stats')om.rutime.model('forecast').gridsearch(X,Y)#usetheRESTAPItostoreadretrievedata,rupredictiosrequests.put('/v1/dataset/stats',jso={...})requests.get('/v1/dataset/stats?sales__gte=100')requests.put('/v1/model/forecast',jso={...})UseCasesomega|mlcurretlysupportsscikit-lear,KerasadTesorflowoutofthebox.Needtodeployamodelfromaotherframework?Opeaissueathttps://github.com/omegaml/omegaml/issuesordropusalieatsupport@omegaml.io
MachieLearigDeploymetdeploymodelstoproductiowithasiglelieofcodeserveadusemodelsordatasetsfromaRESTAPIDataScieceCollaboratiogetafullyitegrateddatascieceworkplacewithimiuteseasilysharemodels,data,jupyterotebooksadreportswithyourcollaboratorsCetralizedData&Computeclusterperformout-of-corecomputatiosoapure-pythoorApacheSparkcomputeclusterhaveasharedNoSQLdatabase(MogoDB),outofthebox,workiglikeaPadasdataframeuseacomputeclustertotraiyourmodelswithoadditioalsetupScalabilityadExtesibilityscaleyourdatascieceworkfromyourlaptoptoteamtoproductiowithocodechagesitegrateaymachielearigframeworkorthirdpartydatascieceplatformwithacommoAPITowardsDataSciecerecetlypublishedaarticleoomega|ml:https://towardsdatasciece.com/omega-ml-deployig-data-machie-learig-pipelies-the-easy-way-a3d281569666
Iadditioomega|mlprovidesaeasy-to-useextesiosAPItosupportaykidofmodels,computecluster,databaseaddatasource.
EterpriseEditio
https://omegaml.io
omega|mlEterpriseEditioprovidessecurityoeveryleveladisreadymadeforKuberetesdeploymet.Itislicesedseparatelyforo-premise,privateorhybridcloud.Sigupathttps://omegaml.io
评论