{"id":197795,"date":"2023-06-07T08:33:43","date_gmt":"2023-06-07T08:33:43","guid":{"rendered":"https:\/\/datascientest.com\/?p=197795"},"modified":"2023-11-09T13:52:09","modified_gmt":"2023-11-09T13:52:09","slug":"cuml-tout-savoir","status":"publish","type":"post","link":"https:\/\/datascientest.com\/cuml-tout-savoir","title":{"rendered":"cuML : D\u00e9finition et utilisation en Machine Learning"},"content":{"rendered":"\t\t<div data-elementor-type=\"wp-post\" data-elementor-id=\"197795\" class=\"elementor elementor-197795\" data-elementor-post-type=\"post\">\n\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-d0d0e3f e-flex e-con-boxed e-con\" data-id=\"d0d0e3f\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-7c21839 elementor-widget elementor-widget-text-editor\" data-id=\"7c21839\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.15.0 - 02-08-2023 *\/\n.elementor-widget-text-editor.elementor-drop-cap-view-stacked .elementor-drop-cap{background-color:#69727d;color:#fff}.elementor-widget-text-editor.elementor-drop-cap-view-framed .elementor-drop-cap{color:#69727d;border:3px solid;background-color:transparent}.elementor-widget-text-editor:not(.elementor-drop-cap-view-default) .elementor-drop-cap{margin-top:8px}.elementor-widget-text-editor:not(.elementor-drop-cap-view-default) .elementor-drop-cap-letter{width:1em;height:1em}.elementor-widget-text-editor .elementor-drop-cap{float:left;text-align:center;line-height:1;font-size:50px}.elementor-widget-text-editor .elementor-drop-cap-letter{display:inline-block}<\/style>\t\t\t\t<p><span style=\"font-weight: 400;\">Lorsque nous faisons face \u00e0 des datasets de grandes dimensions avec \u00e9norm\u00e9ment d\u2019observations et \u00e0 des mod\u00e8les de Machine Learning complexes avec beaucoup d\u2019hyperparam\u00e8tres, l\u2019\u00e9tape d&rsquo;entra\u00eenement peut prendre plusieurs heures. Dans ce cas, une relation inverse entre le nombre de donn\u00e9es et la performance du mod\u00e8le se met en place. Alors pour r\u00e9duire ce temps perdu \u00e0 attendre l&rsquo;ex\u00e9cution des algorithmes et \u00e9viter cette diminution de performance, <\/span><b>cuML semble \u00eatre la solution !<\/b><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-70f9dad e-flex e-con-boxed e-con\" data-id=\"70f9dad\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-51b0657 elementor-widget elementor-widget-heading\" data-id=\"51b0657\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.15.0 - 02-08-2023 *\/\n.elementor-heading-title{padding:0;margin:0;line-height:1}.elementor-widget-heading .elementor-heading-title[class*=elementor-size-]>a{color:inherit;font-size:inherit;line-height:inherit}.elementor-widget-heading .elementor-heading-title.elementor-size-small{font-size:15px}.elementor-widget-heading .elementor-heading-title.elementor-size-medium{font-size:19px}.elementor-widget-heading .elementor-heading-title.elementor-size-large{font-size:29px}.elementor-widget-heading .elementor-heading-title.elementor-size-xl{font-size:39px}.elementor-widget-heading .elementor-heading-title.elementor-size-xxl{font-size:59px}<\/style><h3 class=\"elementor-heading-title elementor-size-default\">Qu'est ce que cuML ?<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-c0a8e38 elementor-widget elementor-widget-text-editor\" data-id=\"c0a8e38\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p><span style=\"font-weight: 400;\">cuML est une <strong>API OpenSource d\u2019algorithmes<\/strong>. Elle fait partie de la suite de <strong>packages<\/strong> <\/span><b>RAPIDS<\/b><span style=\"font-weight: 400;\">, dont la caract\u00e9ristique principale est d\u2019utiliser des GPU (Graphical Processing Unit) pour toutes les structures d\u2019algorithmes, et pas seulement celles de r\u00e9seaux de neurones comme habituellement. Plus pr\u00e9cis\u00e9ment, cuML utilise un <\/span><a href=\"https:\/\/datascientest.com\/gpu-data-science\">GPU<\/a> <b>acc\u00e9l\u00e9r\u00e9<\/b><span style=\"font-weight: 400;\"> pouvant multiplier les performances d\u2019un CPU (Central Processing Unit) classique par 4 pour des <a href=\"https:\/\/datascientest.com\/3-algorithmes-de-machine-learning-utiles-pour-votre-business\">algorithmes de Machine Learning<\/a> simples et par 1000 pour des algorithmes plus complexes.\u00a0<\/span><\/p><p><span style=\"font-weight: 400;\">Gr\u00e2ce \u00e0 cela, c\u2019est une r\u00e9elle alternative \u00e0 Scikit-Learn. En effet Scikit-Learn est une biblioth\u00e8que Python certes compl\u00e8te et tr\u00e8s utilis\u00e9e en Data Science, mais peu efficace en termes de rapidit\u00e9 pour des Big Data.<\/span><\/p><p><span style=\"font-weight: 400;\">Prenons un exemple pour comparer le temps d&rsquo;ex\u00e9cution des deux librairies. Nous avons multipli\u00e9 par 8000 le dataset Titanic disponible sur <a href=\"https:\/\/datascientest.com\/kaggle-tout-ce-quil-a-savoir-sur-cette-plateforme\">Kaggle<\/a> afin d\u2019obtenir un dataset de plus de sept million d\u2019observations. Nous souhaitons faire de la classification binaire pour pr\u00e9dire si un passager du Titanic survit ou non. Avec <a href=\"https:\/\/datascientest.com\/tout-savoir-sur-scikit-learn\">Scikit-Learn<\/a>, le temps d\u2019entra\u00eenement d\u2019un mod\u00e8le Random Forest sur le dataframe agrandi est de 143 secondes. Avec cuML, le temps d\u2019entra\u00eenement d\u2019un mod\u00e8le Random Forest sur ce m\u00eame dataframe est de de seulement 1.6 seconde. Cela prouve \u00e0 quel point cuML permet de <strong>travailler beaucoup plus efficacement et rapidement sur un projet de Data Science<\/strong> !<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-e51ce04 elementor-widget elementor-widget-image\" data-id=\"e51ce04\" data-element_type=\"widget\" data-widget_type=\"image.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.15.0 - 02-08-2023 *\/\n.elementor-widget-image{text-align:center}.elementor-widget-image a{display:inline-block}.elementor-widget-image a img[src$=\".svg\"]{width:48px}.elementor-widget-image img{vertical-align:middle;display:inline-block}<\/style>\t\t\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"767\" height=\"524\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image1-5.png\" class=\"attachment-full size-full wp-image-197801\" alt=\"\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image1-5.png 767w, https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image1-5-300x205.png 300w\" sizes=\"(max-width: 767px) 100vw, 767px\" \/>\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-1881449 elementor-align-center button-principale elementor-widget elementor-widget-button\" data-id=\"1881449\" data-element_type=\"widget\" data-widget_type=\"button.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t<div class=\"elementor-button-wrapper\">\n\t\t\t<a class=\"elementor-button elementor-button-link elementor-size-sm elementor-animation-grow\" href=\"https:\/\/datascientest.com\/formation-machine-learning-engineer\">\n\t\t\t\t\t\t<span class=\"elementor-button-content-wrapper\">\n\t\t\t\t\t\t<span class=\"elementor-button-text\">Apprendre \u00e0 utiliser cuML<\/span>\n\t\t<\/span>\n\t\t\t\t\t<\/a>\n\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-bec2d3b e-flex e-con-boxed e-con\" data-id=\"bec2d3b\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-fc479e2 elementor-widget elementor-widget-heading\" data-id=\"fc479e2\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h3 class=\"elementor-heading-title elementor-size-default\">Mais comment utiliser cuML ?<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-e20317b elementor-widget elementor-widget-text-editor\" data-id=\"e20317b\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<ol><li style=\"font-weight: 400;\" aria-level=\"1\"><b>Installation : <\/b><span style=\"font-weight: 400;\">cuML fait partie de la suite d\u2019API RAPIDS. Il faut ainsi l\u2019installer en passant par le\u00a0<a href=\"https:\/\/docs.rapids.ai\/install#rapids-release-selector\" target=\"_blank\" rel=\"noopener\">RAPIDS Release Selector<\/a>, accessible via ce lien<\/span><span style=\"font-weight: 400;\">, et en cochant cuML et les options compatibles avec la machine utilis\u00e9e.<\/span><\/li><li aria-level=\"1\"><b style=\"font-weight: 400;\">Utilisation : <\/b><span style=\"font-weight: 400;\">L\u2019API de cuML est tr\u00e8s proche de celle de Scikit-Learn, ce qui la rend facile d\u2019utilisation. En effet, elle contient des <\/span><b>outils statistiques<\/b><span style=\"font-weight: 400;\">, des <\/span><b>Scaler de preprocessing<\/b><span style=\"font-weight: 400;\"> et des <\/span><b>m\u00e9thodes de Tuning d\u2019hyperparam\u00e8tres<\/b><span style=\"font-weight: 400;\"> d\u00e9finis de la m\u00eame fa\u00e7on. Pour la mod\u00e9lisation, il suffit d\u2019emporter le mod\u00e8le choisi depuis la librairie cuML, de lui <\/span><b>passer les hyperparam\u00e8tres en argument<\/b><span style=\"font-weight: 400;\">, puis d\u2019utiliser la m\u00e9thode .fit() pour l&rsquo;entra\u00eener.<br \/><\/span>Poursuivons avec le probl\u00e8me du Titanic. Pour entra\u00eener le mod\u00e8le Random Forest avec 200 arbres, nous utilisons le code suivant :\u00a0<\/li><\/ol>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-c75c36e elementor-widget elementor-widget-shortcode\" data-id=\"c75c36e\" data-element_type=\"widget\" data-widget_type=\"shortcode.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t<div class=\"elementor-shortcode\"><script src=\"https:\/\/gist.github.com\/manongeorget\/2c275dbb2351b36e7b13ae1aa18bcfff.js\"><\/script><\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-89e67d8 elementor-widget elementor-widget-text-editor\" data-id=\"89e67d8\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p style=\"padding-left: 40px;\"><span style=\"font-weight: 400;\">Ce code ressemble bien \u00e0 celui que nous aurions \u00e9crit en utilisant Scikit-Learn :<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-b2ee230 elementor-widget elementor-widget-shortcode\" data-id=\"b2ee230\" data-element_type=\"widget\" data-widget_type=\"shortcode.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t<div class=\"elementor-shortcode\"><script src=\"https:\/\/gist.github.com\/manongeorget\/622e140981e08ab8cb7de839c0f3d651.js\"><\/script><\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-051cd2c e-flex e-con-boxed e-con\" data-id=\"051cd2c\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-8672176 elementor-widget elementor-widget-heading\" data-id=\"8672176\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h3 class=\"elementor-heading-title elementor-size-default\">Quelles sont les performances de pr\u00e9dictions ?<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-76342b9 elementor-widget elementor-widget-text-editor\" data-id=\"76342b9\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p><span style=\"font-weight: 400;\">cuML est donc <strong>simple d\u2019utilisation pour les connaisseurs de Scikit-Learn<\/strong> et surpasse m\u00eame cette librairie d\u2019un point de vue de rapidit\u00e9.\u00a0<\/span><\/p><p><span style=\"font-weight: 400;\">En termes de performances, elle semble tout aussi efficace. Que ce soit en r\u00e9gression, en classification ou en clustering, il a \u00e9t\u00e9 prouv\u00e9 que les <strong>mod\u00e8les de cuML<\/strong> obtiennent g\u00e9n\u00e9ralement d\u2019aussi bons r\u00e9sultats de pr\u00e9dictions que ceux de Scikit-Learn. Il est m\u00eame possible de tester un ensemble d\u2019hyperparam\u00e8tres pour obtenir un mod\u00e8le optimal, de faire de la feature selection et d\u2019interpr\u00e9ter les mod\u00e8les.\u00a0<\/span><\/p><p><span style=\"font-weight: 400;\">Ainsi cuML semble \u00eatre la solution id\u00e9ale pour <\/span><b>traiter des grosses bases de donn\u00e9es<\/b><span style=\"font-weight: 400;\"> ou des donn\u00e9es qui continuent d\u2019\u00e9voluer. \u00c0 la diff\u00e9rence de Scikit-Learn, peu comp\u00e9titif en termes de temps, et de <a href=\"https:\/\/datascientest.com\/apache-spark\">Spark<\/a>, bas\u00e9 sur un CPU classique, cuML utilise les fonctionnalit\u00e9s d\u2019un<\/span><b> GPU acc\u00e9l\u00e9r\u00e9<\/b><span style=\"font-weight: 400;\">. Cela lui permet ainsi d\u2019optimiser des mod\u00e8les complexes en un temps r\u00e9duit !<\/span><\/p><p><span style=\"font-weight: 400;\">Maintenant que vous savez tout sur cuML, vous souhaitez peut-\u00eatre le mettre en pratique dans vos projets data. Pour ce faire, <\/span><b>DataScientest vous invite \u00e0 d\u00e9couvrir ses formations data.<\/b><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-c0ca254 elementor-align-center button-principale elementor-widget elementor-widget-button\" data-id=\"c0ca254\" data-element_type=\"widget\" data-widget_type=\"button.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t<div class=\"elementor-button-wrapper\">\n\t\t\t<a class=\"elementor-button elementor-button-link elementor-size-sm elementor-animation-grow\" href=\"https:\/\/datascientest.com\/formation-machine-learning-engineer\">\n\t\t\t\t\t\t<span class=\"elementor-button-content-wrapper\">\n\t\t\t\t\t\t<span class=\"elementor-button-text\">D\u00e9couvrir la formation Machine Learning Engineer<\/span>\n\t\t<\/span>\n\t\t\t\t\t<\/a>\n\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-02c49a9 e-flex e-con-boxed e-con\" data-id=\"02c49a9\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-397bd21 elementor-widget elementor-widget-html\" data-id=\"397bd21\" data-element_type=\"widget\" data-widget_type=\"html.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<script type=\"application\/ld+json\">\r\n{\r\n  \"@context\": \"https:\/\/schema.org\",\r\n  \"@type\": \"FAQPage\",\r\n  \"mainEntity\": [{\r\n    \"@type\": \"Question\",\r\n    \"name\": \"Qu'est ce que cuML ?\",\r\n    \"acceptedAnswer\": {\r\n      \"@type\": \"Answer\",\r\n      \"text\": \"cuML est une API OpenSource d\u2019algorithmes. Elle fait partie de la suite de packages RAPIDS, dont la caract\u00e9ristique principale est d\u2019utiliser des GPU (Graphical Processing Unit) pour toutes les structures d\u2019algorithmes, et pas seulement celles de r\u00e9seaux de neurones comme habituellement. Plus pr\u00e9cis\u00e9ment, cuML utilise un GPU acc\u00e9l\u00e9r\u00e9 pouvant multiplier les performances d\u2019un CPU (Central Processing Unit) classique par 4 pour des algorithmes de Machine Learning simples et par 1000 pour des algorithmes plus complexes.\"\r\n    }\r\n  },{\r\n    \"@type\": \"Question\",\r\n    \"name\": \"Mais comment utiliser cuML ?\",\r\n    \"acceptedAnswer\": {\r\n      \"@type\": \"Answer\",\r\n      \"text\": \"Les principaux modules et classes utilis\u00e9s sont torch.nn, torch.optim, Dataset et Dataloader. L\u2019informatique quantique est l\u2019une de ses applications s\u00e9duisantes. D\u2019autre part, PyTorch Lightning est un framework open source qui remplace et utilise PyTorch lorsque la complexit\u00e9 du mod\u00e8le augmente et dissocie la recherche de l\u2019ing\u00e9nierie pour une plus op\u00e9rationnalit\u00e9. Il a \u00e9t\u00e9 adopt\u00e9 en 2019 lors de la Conf\u00e9rence: atelier sur les syst\u00e8mes de traitement de l\u2019information neuronale. Utilis\u00e9 aussi bien pour le flux de travail de recherche que pour le pipeline de production, il peut se d\u00e9ployer en quelques jours au lieu de plusieurs mois et \u00e9viter beaucoup de d\u00e9sordre et de code boilerplate (code ind\u00e9pendant de l\u2019exp\u00e9rimentation ou du training).\"\r\n    }\r\n  }]\r\n}\r\n<\/script>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"<p>Lorsque nous faisons face \u00e0 des datasets de grandes dimensions avec \u00e9norm\u00e9ment d\u2019observations et \u00e0 des mod\u00e8les de Machine Learning complexes avec beaucoup d\u2019hyperparam\u00e8tres, l\u2019\u00e9tape d&rsquo;entra\u00eenement peut prendre plusieurs heures. Dans ce cas, une relation inverse entre le nombre de donn\u00e9es et la performance du mod\u00e8le se met en place. Alors pour r\u00e9duire ce temps [&hellip;]<\/p>\n","protected":false},"author":77,"featured_media":197797,"comment_status":"open","ping_status":"open","sticky":false,"template":"elementor_theme","format":"standard","meta":{"content-type":"","inline_featured_image":false},"categories":[2497],"tags":[],"yoast_head":"<!-- This site is optimized with the Yoast SEO Premium plugin v21.7 (Yoast SEO v21.7) - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>cuML : D\u00e9finition et utilisation en Machine Learning | DataScientest<\/title>\n<meta name=\"description\" content=\"Pour r\u00e9duire ce temps perdu \u00e0 attendre l&#039;ex\u00e9cution des algorithmes et \u00e9viter cette diminution de performance, cuML semble \u00eatre la solution. Les d\u00e9tails dans cet article.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/datascientest.com\/cuml-tout-savoir\" \/>\n<meta property=\"og:locale\" content=\"fr_FR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"cuML : D\u00e9finition et utilisation en Machine Learning\" \/>\n<meta property=\"og:url\" content=\"https:\/\/datascientest.com\/cuml-tout-savoir\" \/>\n<meta property=\"og:site_name\" content=\"Formation Data Science | DataScientest.com\" \/>\n<meta property=\"article:published_time\" content=\"2023-06-07T08:33:43+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2023-11-09T13:52:09+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/Illu_BLOG__cuML.png\" \/>\n\t<meta property=\"og:image:width\" content=\"1025\" \/>\n\t<meta property=\"og:image:height\" content=\"563\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"author\" content=\"Raphael Kassel\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u00c9crit par\" \/>\n\t<meta name=\"twitter:data1\" content=\"Raphael Kassel\" \/>\n\t<meta name=\"twitter:label2\" content=\"Dur\u00e9e de lecture estim\u00e9e\" \/>\n\t<meta name=\"twitter:data2\" content=\"4 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/datascientest.com\/cuml-tout-savoir\",\"url\":\"https:\/\/datascientest.com\/cuml-tout-savoir\",\"name\":\"cuML : D\u00e9finition et utilisation en Machine Learning | DataScientest\",\"isPartOf\":{\"@id\":\"https:\/\/datascientest.com\/#website\"},\"datePublished\":\"2023-06-07T08:33:43+00:00\",\"dateModified\":\"2023-11-09T13:52:09+00:00\",\"author\":{\"@id\":\"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1\"},\"description\":\"Pour r\u00e9duire ce temps perdu \u00e0 attendre l'ex\u00e9cution des algorithmes et \u00e9viter cette diminution de performance, cuML semble \u00eatre la solution. Les d\u00e9tails dans cet article.\",\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/datascientest.com\/cuml-tout-savoir\"]}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/datascientest.com\/#website\",\"url\":\"https:\/\/datascientest.com\/\",\"name\":\"Formation Data Science | DataScientest.com\",\"description\":\"Leader de la formation en Data Science\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/datascientest.com\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"fr-FR\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1\",\"name\":\"Raphael Kassel\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\/\/datascientest.com\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/11\/raph.jpeg\",\"contentUrl\":\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/11\/raph.jpeg\",\"caption\":\"Raphael Kassel\"},\"description\":\"Raphael Kassel est le Head of Data Programs chez DataScientest. Il supervise une \u00e9quipe Data, impl\u00e9mente de nouveaux parcours comme \\\"Chef de projet IA\\\" et \\\"Data Product Manager\\\", et contribue \u00e0 la modernisation des op\u00e9rations du groupe. En collaboration \u00e9troite avec l'\u00e9quipe de d\u00e9veloppement, il adapte l'accompagnement acad\u00e9mique aux \u00e9volutions de l'\u00e9quipe data et \u00e0 l'internationalisation.\",\"sameAs\":[\"https:\/\/www.linkedin.com\/in\/raphaelkassel\/\",\"Data Science\"],\"url\":\"https:\/\/datascientest.com\/author\/raphaelkassel\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"cuML : D\u00e9finition et utilisation en Machine Learning | DataScientest","description":"Pour r\u00e9duire ce temps perdu \u00e0 attendre l'ex\u00e9cution des algorithmes et \u00e9viter cette diminution de performance, cuML semble \u00eatre la solution. Les d\u00e9tails dans cet article.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/datascientest.com\/cuml-tout-savoir","og_locale":"fr_FR","og_type":"article","og_title":"cuML : D\u00e9finition et utilisation en Machine Learning","og_url":"https:\/\/datascientest.com\/cuml-tout-savoir","og_site_name":"Formation Data Science | DataScientest.com","article_published_time":"2023-06-07T08:33:43+00:00","article_modified_time":"2023-11-09T13:52:09+00:00","og_image":[{"width":1025,"height":563,"url":"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/Illu_BLOG__cuML.png","type":"image\/png"}],"author":"Raphael Kassel","twitter_card":"summary_large_image","twitter_misc":{"\u00c9crit par":"Raphael Kassel","Dur\u00e9e de lecture estim\u00e9e":"4 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/datascientest.com\/cuml-tout-savoir","url":"https:\/\/datascientest.com\/cuml-tout-savoir","name":"cuML : D\u00e9finition et utilisation en Machine Learning | DataScientest","isPartOf":{"@id":"https:\/\/datascientest.com\/#website"},"datePublished":"2023-06-07T08:33:43+00:00","dateModified":"2023-11-09T13:52:09+00:00","author":{"@id":"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1"},"description":"Pour r\u00e9duire ce temps perdu \u00e0 attendre l'ex\u00e9cution des algorithmes et \u00e9viter cette diminution de performance, cuML semble \u00eatre la solution. Les d\u00e9tails dans cet article.","inLanguage":"fr-FR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/datascientest.com\/cuml-tout-savoir"]}]},{"@type":"WebSite","@id":"https:\/\/datascientest.com\/#website","url":"https:\/\/datascientest.com\/","name":"Formation Data Science | DataScientest.com","description":"Leader de la formation en Data Science","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/datascientest.com\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"fr-FR"},{"@type":"Person","@id":"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1","name":"Raphael Kassel","image":{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/datascientest.com\/#\/schema\/person\/image\/","url":"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/11\/raph.jpeg","contentUrl":"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/11\/raph.jpeg","caption":"Raphael Kassel"},"description":"Raphael Kassel est le Head of Data Programs chez DataScientest. Il supervise une \u00e9quipe Data, impl\u00e9mente de nouveaux parcours comme \"Chef de projet IA\" et \"Data Product Manager\", et contribue \u00e0 la modernisation des op\u00e9rations du groupe. En collaboration \u00e9troite avec l'\u00e9quipe de d\u00e9veloppement, il adapte l'accompagnement acad\u00e9mique aux \u00e9volutions de l'\u00e9quipe data et \u00e0 l'internationalisation.","sameAs":["https:\/\/www.linkedin.com\/in\/raphaelkassel\/","Data Science"],"url":"https:\/\/datascientest.com\/author\/raphaelkassel"}]}},"_links":{"self":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/posts\/197795"}],"collection":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/users\/77"}],"replies":[{"embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/comments?post=197795"}],"version-history":[{"count":0,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/posts\/197795\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/media\/197797"}],"wp:attachment":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/media?parent=197795"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/categories?post=197795"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/tags?post=197795"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}