{"id":197623,"date":"2023-06-07T08:18:34","date_gmt":"2023-06-07T08:18:34","guid":{"rendered":"https:\/\/datascientest.com\/?p=197623"},"modified":"2023-11-09T13:52:12","modified_gmt":"2023-11-09T13:52:12","slug":"hugging-face-transformers-tout-savoir","status":"publish","type":"post","link":"https:\/\/datascientest.com\/hugging-face-transformers-tout-savoir","title":{"rendered":"Hugging Face Transformers : l\u2019API au service du NLP"},"content":{"rendered":"\t\t<div data-elementor-type=\"wp-post\" data-elementor-id=\"197623\" class=\"elementor elementor-197623\" data-elementor-post-type=\"post\">\n\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-b313724 e-flex e-con-boxed e-con\" data-id=\"b313724\" 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-da86fd8 elementor-widget elementor-widget-heading\" data-id=\"da86fd8\" 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><h2 class=\"elementor-heading-title elementor-size-default\">Depuis des ann\u00e9es, l\u2019IA suscite un int\u00e9r\u00eat grandissant avec les performances obtenues gr\u00e2ce \u00e0 l\u2019apprentissage automatique (<a href=\"https:\/\/datascientest.com\/machine-learning-tout-savoir\">Machine Learning<\/a>) et, plus particuli\u00e8rement, avec l\u2019apprentissage profond (<a href=\"https:\/\/datascientest.com\/deep-learning-definition\">Deep Learning<\/a>).<\/h2>\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-a1ae41c e-flex e-con-boxed e-con\" data-id=\"a1ae41c\" 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-09b2c0e elementor-widget elementor-widget-text-editor\" data-id=\"09b2c0e\" 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;\">Ces technologies ont jou\u00e9 un r\u00f4le d\u00e9cisif dans l\u2019\u00e9volution du traitement du langage naturel (<a href=\"https:\/\/datascientest.com\/formation-deep-learning\/natural-language-processing\">NLP<\/a>). Elles ont contribu\u00e9 \u00e0 l&rsquo;<\/span><b>automatisation des t\u00e2ches de NLP<\/b><span style=\"font-weight: 400;\">les plus courantes \u00e0 l\u2019instar de la classification des mots et des phrases, la g\u00e9n\u00e9ration de texte, l\u2019extraction de r\u00e9ponse, et la reconnaissance de la parole et de la vision.<\/span><\/p><p><span style=\"font-weight: 400;\">Aujourd&rsquo;hui, il existe plusieurs <\/span><b>biblioth\u00e8ques open source<\/b><span style=\"font-weight: 400;\"> contenant des mod\u00e8les pr\u00e9-entra\u00een\u00e9s permettant aux Data Scientists et aux entreprises d\u2019utiliser des mod\u00e8les existants afin d&rsquo;\u00e9conomiser en co\u00fbts de calcul et de gagner du temps. Hugging Face est l\u2019une d\u2019entre elles.<\/span><\/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-238261b e-flex e-con-boxed e-con\" data-id=\"238261b\" 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-dfc87d2 elementor-widget elementor-widget-heading\" data-id=\"dfc87d2\" 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\">\u00c0 propos de Hugging Face Transformers<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-35aad50 elementor-widget elementor-widget-text-editor\" data-id=\"35aad50\" 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;\">Hugging Face transformers est une <strong>plateforme destin\u00e9e \u00e0 faire progresser l&rsquo;IA<\/strong> en partageant des collections de mod\u00e8les et des ensembles de donn\u00e9es permettant aux utilisateurs de g\u00e9rer, partager et d\u00e9velopper leurs mod\u00e8les d\u2019IA. Il s\u2019agit d\u2019une<strong> librairie d\u2019apprentissage en profondeur Open Source <\/strong><\/span><span style=\"font-weight: 400;\"><strong>bas\u00e9e sur<\/strong> <a href=\"https:\/\/datascientest.com\/python-tout-savoir\">Python<\/a> et<\/span><span style=\"font-weight: 400;\">offrant une grande vari\u00e9t\u00e9 de mod\u00e8les pr\u00e9-entra\u00een\u00e9s pour diff\u00e9rents types de t\u00e2ches de NLP. Elle pr\u00e9sente une <a href=\"https:\/\/datascientest.com\/api-tout-savoir\">API<\/a> \u00e0 utiliser dans de nombreuses architectures de transformateurs pr\u00e9-entra\u00een\u00e9es, telles que <strong>BERT, GPT-2, DistilBERT, BART <\/strong>ou <strong>GPT-3<\/strong> permettant d\u2019obtenir de tr\u00e8s bons r\u00e9sultats sur une vari\u00e9t\u00e9 de t\u00e2ches NLP comme la classification de texte, la d\u00e9tection d&rsquo;entit\u00e9s nomm\u00e9es, la g\u00e9n\u00e9ration de texte, etc.<\/span><\/p><p><span style=\"font-weight: 400;\">Les transformateurs pr\u00e9-entra\u00een\u00e9es de Hugging Face peuvent \u00eatre regroup\u00e9s en trois cat\u00e9gories :<\/span><\/p><ul><li><b>Les transformers autor\u00e9gressifs <\/b><span style=\"font-weight: 400;\">(architecture de type GPT) : concentr\u00e9s<\/span><span style=\"font-weight: 400;\"> sur la pr\u00e9diction du prochain mot dans la phrase, ils sont adapt\u00e9s pour les t\u00e2ches g\u00e9n\u00e9ratives telles que la g\u00e9n\u00e9ration de texte.<\/span><\/li><li><b>Les transformers auto-encodeurs <\/b><span style=\"font-weight: 400;\">(architecture de type BERT) : <\/span><span style=\"font-weight: 400;\">plus adapt\u00e9s pour des t\u00e2ches qui requi\u00e8rent une compr\u00e9hension compl\u00e8te de l\u2019entr\u00e9e. Par exemple la classification de phrases, la reconnaissance d\u2019entit\u00e9s nomm\u00e9es et plus g\u00e9n\u00e9ralement la classification de mots.<\/span><\/li><li><b>Les transformers s\u00e9quence-\u00e0-s\u00e9quence <\/b><span style=\"font-weight: 400;\">(architecture de type BART\/T5) : <\/span><span style=\"font-weight: 400;\">les plus adapt\u00e9s pour les t\u00e2ches li\u00e9es \u00e0 la g\u00e9n\u00e9ration de nouvelles phrases en fonction d\u2019une entr\u00e9e donn\u00e9e, comme le r\u00e9sum\u00e9 de texte, la traduction ou la g\u00e9n\u00e9ration de questions-r\u00e9ponses.<\/span><\/li><\/ul>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-0063675 elementor-widget elementor-widget-text-editor\" data-id=\"0063675\" 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;\">En plus des mod\u00e8les officiels pr\u00e9-entra\u00een\u00e9s, il existe des centaines de mod\u00e8les de transformateurs de phrases (sentence-transformer) sur le Hugging Face Hub. \u00a0 \u00a0 \u00a0 <\/span><\/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-104c5c9 e-flex e-con-boxed e-con\" data-id=\"104c5c9\" 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-31238a8 elementor-widget elementor-widget-heading\" data-id=\"31238a8\" 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\">Quel est l'int\u00e9r\u00eat de ces Hugging Face Transformers  ? <\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-366e535 elementor-widget elementor-widget-text-editor\" data-id=\"366e535\" 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;\">L&rsquo;objectif des Transformers est de <strong>faciliter le d\u00e9veloppement des projets d&rsquo;apprentissage automatique<\/strong> en fournissant des APIs et des outils pour t\u00e9l\u00e9charger et entra\u00eener des mod\u00e8les pr\u00e9-entra\u00een\u00e9s. L&rsquo;utilisation de ces mod\u00e8les peut r\u00e9duire les co\u00fbts de calcul, et faire \u00e9conomiser le temps et les ressources n\u00e9cessaires pour former un mod\u00e8le \u00e0 partir de z\u00e9ro. Ils <\/span><span style=\"font-weight: 400;\">sont rapidement devenus l&rsquo;architecture dominante pour obtenir des <strong>r\u00e9sultats fiables sur une vari\u00e9t\u00e9 de t\u00e2ches NLP<\/strong>.<\/span><\/p><p><span style=\"font-weight: 400;\">Ces mod\u00e8les prennent en charge des t\u00e2ches courantes dans diff\u00e9rentes modalit\u00e9s :<\/span><\/p><ul><li style=\"font-weight: 400;\" aria-level=\"1\"><b>Texte : <\/b><span style=\"font-weight: 400;\">Pour tout ce qui touche \u00e0 la classification, l&rsquo;extraction d&rsquo;informations, la r\u00e9ponse \u00e0 des questions, la g\u00e9n\u00e9ration, la cr\u00e9ation et la traduction de texte dans plus de 100 langues diff\u00e9rentes.<\/span><\/li><li style=\"font-weight: 400;\" aria-level=\"1\"><b>Discours<\/b><span style=\"font-weight: 400;\"> : Pour des t\u00e2ches telles que la classification audio des objets et la reconnaissance vocale.<\/span><\/li><li style=\"font-weight: 400;\" aria-level=\"1\"><b>Vision : <\/b><span style=\"font-weight: 400;\">Pour la d\u00e9tection d&rsquo;objets, la classification d&rsquo;images et la segmentation.<\/span><\/li><li style=\"font-weight: 400;\" aria-level=\"1\"><b>Donn\u00e9es tabulaires<\/b><span style=\"font-weight: 400;\"> : Pour les probl\u00e8mes de r\u00e9gression et de classification.<\/span><\/li><\/ul>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-c54402e elementor-widget elementor-widget-text-editor\" data-id=\"c54402e\" 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;\">D\u2019autre part, les Transformers prennent en charge l&rsquo;interop\u00e9rabilit\u00e9 des frameworks entre <a href=\"https:\/\/datascientest.com\/pytorch-tout-savoir\">PyTorch<\/a>, <a href=\"https:\/\/datascientest.com\/formation-tensorflow\">TensorFlow<\/a> et <strong>JAX<\/strong>. Cela permet d&rsquo;utiliser un framework diff\u00e9rent \u00e0 chaque \u00e9tape de la vie d&rsquo;un mod\u00e8le : former un mod\u00e8le en trois lignes de code dans un framework, et le charger pour l&rsquo;inf\u00e9rence dans un autre.<\/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-4e5fd48 elementor-align-center button-principale elementor-widget elementor-widget-button\" data-id=\"4e5fd48\" 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-deep-learning\">\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 Hugging Face Transformers<\/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-1ede827 e-flex e-con-boxed e-con\" data-id=\"1ede827\" 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-8749e38 elementor-widget elementor-widget-heading\" data-id=\"8749e38\" 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\">Mise en place des Transformers<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-b47cada elementor-widget elementor-widget-text-editor\" data-id=\"b47cada\" 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;\">Dans cette partie, nous allons mettre en pratique les Transformers en d\u00e9couvrant l&rsquo;outil <\/span><i><span style=\"font-weight: 400;\">pipeline<\/span><\/i><span style=\"font-weight: 400;\">, puis nous montrerons comment cet outil facilite l&rsquo;impl\u00e9mentation des t\u00e2ches de NLP et de <a href=\"https:\/\/datascientest.com\/formation-deep-learning\/formation-computer-vision\">computer vision<\/a>.<\/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-bce9865 elementor-widget elementor-widget-heading\" data-id=\"bce9865\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h4 class=\"elementor-heading-title elementor-size-default\">Les Pipelines<\/h4>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-ef63aa8 elementor-widget elementor-widget-text-editor\" data-id=\"ef63aa8\" 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;\">Les pipelines sont <strong>un moyen simple et efficace<\/strong> pour extraire une grande partie du code complexe de la biblioth\u00e8que fournissant une API pour<strong> effectuer des inf\u00e9rences sur une vari\u00e9t\u00e9 de t\u00e2ches<\/strong>, notamment la reconnaissance d\u2019entit\u00e9s nomm\u00e9es, l\u2019analyse des sentiments, la r\u00e9ponse aux questions, etc.<\/span><\/p><p><span style=\"font-weight: 400;\">Ils sont utilis\u00e9s pour encapsuler le processus global de chaque t\u00e2che de NLP comme le nettoyage du texte, la tokenisation, l&rsquo;incorporation, etc.<\/span><\/p><p><span style=\"font-weight: 400;\">C\u2019est une fonction simple \u00e0 utiliser : l\u2019instanciation d\u2019un pipeline se fait via la <strong>m\u00e9thode pipeline()<\/strong>. Il suffit de sp\u00e9cifier la t\u00e2che \u00e0 lancer.<\/span><\/p><p><span style=\"font-weight: 400;\">Ci-apr\u00e8s, une liste non-exhaustive des <a href=\"https:\/\/huggingface.co\/transformers\/main_classes\/pipelines.html\">pipelines disponibles<\/a> <\/span><span style=\"font-weight: 400;\">de la biblioth\u00e8que Hugging Face Transformers :<\/span><\/p><ul><li><span style=\"font-weight: 400;\">feature-extraction (pour obtenir la repr\u00e9sentation vectorielle d\u2019un texte)<\/span><\/li><li><span style=\"font-weight: 400;\">fill-mask (pour pr\u00e9dire les mots masqu\u00e9s)<\/span><\/li><li><span style=\"font-weight: 400;\">ner (reconnaissance d\u2019entit\u00e9s nomm\u00e9es)<\/span><\/li><li><span style=\"font-weight: 400;\">question-answering (pour r\u00e9pondre aux questions)<\/span><\/li><li><span style=\"font-weight: 400;\">sentiment-analysis (pour analyser les sentiments)<\/span><\/li><li><span style=\"font-weight: 400;\">summarization (pour r\u00e9sumer un texte)<\/span><\/li><li><span style=\"font-weight: 400;\">text-generation (pour produire un nouveau texte)<\/span><\/li><li><span style=\"font-weight: 400;\">translation (pour traduire un texte)<\/span><\/li><li><span style=\"font-weight: 400;\">zero-shot-classification (pour classer de nouveaux exemples \u00e0 partir de classes in\u00e9dites)<\/span><\/li><\/ul>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-2dcd1cf elementor-widget elementor-widget-heading\" data-id=\"2dcd1cf\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h4 class=\"elementor-heading-title elementor-size-default\">G\u00e9n\u00e9ration de Texte<\/h4>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-f4368d2 elementor-widget elementor-widget-text-editor\" data-id=\"f4368d2\" 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;\">De nos jours, la g\u00e9n\u00e9ration de texte est devenue une tendance avec la publication et l&rsquo;\u00e9volution de <a href=\"https:\/\/datascientest.com\/chatgpt-se-met-a-jour-et-obtient-lacces-a-internet\">ChatGPT<\/a>.\u00a0<\/span><span style=\"font-weight: 400;\">Dans ce contexte, le pipeline nous permet d&rsquo;utiliser des mod\u00e8les avanc\u00e9s de g\u00e9n\u00e9ration de texte en une seule ligne de code.<\/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-b14e2c3 elementor-widget elementor-widget-image\" data-id=\"b14e2c3\" 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=\"688\" height=\"105\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image2-6.png\" class=\"attachment-full size-full wp-image-197638\" alt=\"\" loading=\"lazy\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image2-6.png 688w, https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image2-6-300x46.png 300w\" sizes=\"(max-width: 688px) 100vw, 688px\" \/>\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-d9e66c0 elementor-widget elementor-widget-text-editor\" data-id=\"d9e66c0\" 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;\">Dans la ligne ci-dessus, nous sp\u00e9cifions le cas d&rsquo;utilisation que nous souhaitons (g\u00e9n\u00e9ration de texte), et il va automatiquement t\u00e9l\u00e9charger le mod\u00e8le par d\u00e9faut (dans ce cas GPT-2). Nous pouvons utiliser d&rsquo;autres versions en sp\u00e9cifiant le param\u00e8tre &lsquo;model&rsquo;.<\/span><\/p><p><span style=\"font-weight: 400;\">Maintenant commen\u00e7ons \u00e0 g\u00e9n\u00e9rer le texte en passant une phrase comme un param\u00e8tre.<\/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-beefe0a elementor-widget elementor-widget-image\" data-id=\"beefe0a\" data-element_type=\"widget\" data-widget_type=\"image.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"738\" height=\"66\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image4-5.png\" class=\"attachment-full size-full wp-image-197639\" alt=\"\" loading=\"lazy\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image4-5.png 738w, https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image4-5-300x27.png 300w\" sizes=\"(max-width: 738px) 100vw, 738px\" \/>\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-d99e9fc elementor-widget elementor-widget-text-editor\" data-id=\"d99e9fc\" 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;\">Dans cet exemple, nous avons pos\u00e9 une question au mod\u00e8le \u00a0\u00bb Why artificial intelligence ? \u00ab\u00a0.<\/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-1e01a8e elementor-widget elementor-widget-image\" data-id=\"1e01a8e\" data-element_type=\"widget\" data-widget_type=\"image.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"1051\" height=\"77\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image3-4.png\" class=\"attachment-full size-full wp-image-197640\" alt=\"\" loading=\"lazy\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image3-4.png 1051w, https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image3-4-300x22.png 300w, https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image3-4-1024x75.png 1024w, https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image3-4-768x56.png 768w\" sizes=\"(max-width: 1051px) 100vw, 1051px\" \/>\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-720d514 elementor-widget elementor-widget-text-editor\" data-id=\"720d514\" 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;\">Le mod\u00e8le a g\u00e9n\u00e9r\u00e9 une r\u00e9ponse dans la cl\u00e9 &lsquo;generated_text&rsquo;, et nous pouvons g\u00e9n\u00e9rer une autre r\u00e9ponse en <\/span><b>ex\u00e9cutant le code une autre fois<\/b><span style=\"font-weight: 400;\">. Cet exemple montre que les Transformers nous aident \u00e0 mettre en \u0153uvre un chatbot en un temps r\u00e9duit.<\/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-526fe0c elementor-widget elementor-widget-heading\" data-id=\"526fe0c\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h4 class=\"elementor-heading-title elementor-size-default\">Classification d\u2019Image<\/h4>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-6033007 elementor-widget elementor-widget-text-editor\" data-id=\"6033007\" 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;\">La classification d\u2019images est un probl\u00e8me li\u00e9 \u00e0 l\u2019apprentissage automatique (Machine Learning). Avec <\/span><b>l\u2019API Transformers<\/b>,<span style=\"font-weight: 400;\"> il est simple d\u2019acqu\u00e9rir le mod\u00e8le pour ce probl\u00e8me.<\/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-2a53781 elementor-widget elementor-widget-image\" data-id=\"2a53781\" data-element_type=\"widget\" data-widget_type=\"image.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"816\" height=\"97\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image6-4.png\" class=\"attachment-full size-full wp-image-197641\" alt=\"\" loading=\"lazy\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image6-4.png 816w, https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image6-4-300x36.png 300w, https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image6-4-768x91.png 768w\" sizes=\"(max-width: 816px) 100vw, 816px\" \/>\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-c4c0e5a elementor-widget elementor-widget-text-editor\" data-id=\"c4c0e5a\" 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;\">En une ligne, nous avons acquis le mod\u00e8le Image Classifier. Nous pourrions facilement passer un fichier image de n\u2019importe quelle source pour identifier l\u2019image avec ce mod\u00e8le.<\/span><\/p><p><span style=\"font-weight: 400;\">Par exemple, nous allons passer le lien de l\u2019image au dessous et laisser le mod\u00e8le la classifier.<\/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-a15e6ad elementor-widget elementor-widget-image\" data-id=\"a15e6ad\" data-element_type=\"widget\" data-widget_type=\"image.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"373\" height=\"278\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image5-3.png\" class=\"attachment-full size-full wp-image-197642\" alt=\"\" loading=\"lazy\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image5-3.png 373w, https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image5-3-300x224.png 300w\" sizes=\"(max-width: 373px) 100vw, 373px\" \/>\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-83416a6 elementor-widget elementor-widget-text-editor\" data-id=\"83416a6\" 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;\">La Pipeline nous aide \u00e0 le faire en une seule ligne de code. Le mod\u00e8le propose cinq classifications, chacune associ\u00e9e \u00e0 une probabilit\u00e9. La premi\u00e8re a le plus grand score, nous pouvons donc\u00a0 prendre \u00ab\u00a0voiture de sport\u00a0\u00bb comme pr\u00e9diction principale et c\u2019est en effet juste.<\/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-df453e3 elementor-widget elementor-widget-image\" data-id=\"df453e3\" data-element_type=\"widget\" data-widget_type=\"image.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"800\" height=\"146\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image1-4.png\" class=\"attachment-large size-large wp-image-197643\" alt=\"\" loading=\"lazy\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image1-4.png 940w, https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image1-4-300x55.png 300w, https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/image1-4-768x140.png 768w\" sizes=\"(max-width: 800px) 100vw, 800px\" \/>\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-33d0cbf elementor-widget elementor-widget-text-editor\" data-id=\"33d0cbf\" 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;\">Cet exemple prouve que nous pouvons mettre en \u0153uvre un mod\u00e8le sophistiqu\u00e9 en peu de temps et sans avoir de grandes connaissances en IA.<\/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-9ff9ecd elementor-widget elementor-widget-heading\" data-id=\"9ff9ecd\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h4 class=\"elementor-heading-title elementor-size-default\">Sp\u00e9cification du mod\u00e8le dans un pipeline<\/h4>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-95a4eb5 elementor-widget elementor-widget-text-editor\" data-id=\"95a4eb5\" 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;\">Dans les exemples pr\u00e9c\u00e9dents nous avons utilis\u00e9 le mod\u00e8le par d\u00e9faut pour la t\u00e2che en question, mais il est aussi possible de choisir un mod\u00e8le particulier du <\/span><i><span style=\"font-weight: 400;\">Hub<\/span><\/i><span style=\"font-weight: 400;\"> et l\u2019utiliser dans un pipeline pour une t\u00e2che sp\u00e9cifique.<\/span><\/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-e945e99 e-flex e-con-boxed e-con\" data-id=\"e945e99\" 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-d9da10b elementor-widget elementor-widget-heading\" data-id=\"d9da10b\" 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\">Conclusion<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-1144205 elementor-widget elementor-widget-text-editor\" data-id=\"1144205\" 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;\">Aujourd\u2019hui, Hugging Face Transformers est <\/span><span style=\"font-weight: 400;\">la biblioth\u00e8que logicielle la plus largement adopt\u00e9e, aussi bien par les d\u00e9butants que par les professionnels pour les mod\u00e8les d\u2019apprentissage automatique afin de <strong>traiter les t\u00e2ches NLP<\/strong>.<\/span><span style=\"font-weight: 400;\"> Elle met \u00e0 la disposition de l\u2019utilisateur un mod\u00e8le pr\u00e9-entra\u00een\u00e9 qui \u00e9tait auparavant inaccessible en raison des exigences techniques. <\/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-8569577 elementor-align-center button-principale elementor-widget elementor-widget-button\" data-id=\"8569577\" 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-deep-learning\">\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\u00e9marrer une formation en Deep Learning<\/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-c9a819f e-flex e-con-boxed e-con\" data-id=\"c9a819f\" 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-79a5fcd elementor-widget elementor-widget-html\" data-id=\"79a5fcd\" 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\": \"\u00c0 propos de Hugging Face Transformers\",\r\n    \"acceptedAnswer\": {\r\n      \"@type\": \"Answer\",\r\n      \"text\": \"Hugging Face transformers est une plateforme destin\u00e9e \u00e0 faire progresser l'IA en partageant des collections de mod\u00e8les et des ensembles de donn\u00e9es permettant aux utilisateurs de g\u00e9rer, partager et d\u00e9velopper leurs mod\u00e8les d\u2019IA. Il s\u2019agit d\u2019une librairie d\u2019apprentissage en profondeur Open Source bas\u00e9e sur Python etoffrant une grande vari\u00e9t\u00e9 de mod\u00e8les pr\u00e9-entra\u00een\u00e9s pour diff\u00e9rents types de t\u00e2ches de NLP. Elle pr\u00e9sente une API \u00e0 utiliser dans de nombreuses architectures de transformateurs pr\u00e9-entra\u00een\u00e9es, telles que BERT, GPT-2, DistilBERT, BART ou GPT-3 permettant d\u2019obtenir de tr\u00e8s bons r\u00e9sultats sur une vari\u00e9t\u00e9 de t\u00e2ches NLP comme la classification de texte, la d\u00e9tection d'entit\u00e9s nomm\u00e9es, la g\u00e9n\u00e9ration de texte, etc.\"\r\n    }\r\n  },{\r\n    \"@type\": \"Question\",\r\n    \"name\": \"Quel est l'int\u00e9r\u00eat de ces Hugging Face Transformers  ? \",\r\n    \"acceptedAnswer\": {\r\n      \"@type\": \"Answer\",\r\n      \"text\": \"L'objectif des Transformers est de faciliter le d\u00e9veloppement des projets d'apprentissage automatique en fournissant des APIs et des outils pour t\u00e9l\u00e9charger et entra\u00eener des mod\u00e8les pr\u00e9-entra\u00een\u00e9s. L'utilisation de ces mod\u00e8les peut r\u00e9duire les co\u00fbts de calcul, et faire \u00e9conomiser le temps et les ressources n\u00e9cessaires pour former un mod\u00e8le \u00e0 partir de z\u00e9ro. Ils sont rapidement devenus l'architecture dominante pour obtenir des r\u00e9sultats fiables sur une vari\u00e9t\u00e9 de t\u00e2ches NLP.\"\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>Depuis des ann\u00e9es, l\u2019IA suscite un int\u00e9r\u00eat grandissant avec les performances obtenues gr\u00e2ce \u00e0 l\u2019apprentissage automatique (Machine Learning) et, plus particuli\u00e8rement, avec l\u2019apprentissage profond (Deep Learning). Ces technologies ont jou\u00e9 un r\u00f4le d\u00e9cisif dans l\u2019\u00e9volution du traitement du langage naturel (NLP). Elles ont contribu\u00e9 \u00e0 l&rsquo;automatisation des t\u00e2ches de NLPles plus courantes \u00e0 l\u2019instar de [&hellip;]<\/p>\n","protected":false},"author":77,"featured_media":197625,"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>Hugging Face Transformers : l\u2019API au service du NLP<\/title>\n<meta name=\"description\" content=\"Il existe plusieurs biblioth\u00e8ques open source contenant des mod\u00e8les pr\u00e9-entra\u00een\u00e9s permettant aux Data Scientists et aux entreprises d\u2019utiliser des mod\u00e8les existants afin d&#039;\u00e9conomiser en co\u00fbts de calcul et de gagner du temps. Hugging Face est l\u2019une d\u2019entre elles.\" \/>\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\/hugging-face-transformers-tout-savoir\" \/>\n<meta property=\"og:locale\" content=\"fr_FR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Hugging Face Transformers : l\u2019API au service du NLP\" \/>\n<meta property=\"og:url\" content=\"https:\/\/datascientest.com\/hugging-face-transformers-tout-savoir\" \/>\n<meta property=\"og:site_name\" content=\"Formation Data Science | DataScientest.com\" \/>\n<meta property=\"article:published_time\" content=\"2023-06-07T08:18:34+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2023-11-09T13:52:12+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/Illu_BLOG__hugging_face.png\" \/>\n\t<meta property=\"og:image:width\" content=\"1024\" \/>\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=\"8 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/datascientest.com\/hugging-face-transformers-tout-savoir\",\"url\":\"https:\/\/datascientest.com\/hugging-face-transformers-tout-savoir\",\"name\":\"Hugging Face Transformers : l\u2019API au service du NLP\",\"isPartOf\":{\"@id\":\"https:\/\/datascientest.com\/#website\"},\"datePublished\":\"2023-06-07T08:18:34+00:00\",\"dateModified\":\"2023-11-09T13:52:12+00:00\",\"author\":{\"@id\":\"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1\"},\"description\":\"Il existe plusieurs biblioth\u00e8ques open source contenant des mod\u00e8les pr\u00e9-entra\u00een\u00e9s permettant aux Data Scientists et aux entreprises d\u2019utiliser des mod\u00e8les existants afin d'\u00e9conomiser en co\u00fbts de calcul et de gagner du temps. Hugging Face est l\u2019une d\u2019entre elles.\",\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/datascientest.com\/hugging-face-transformers-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":"Hugging Face Transformers : l\u2019API au service du NLP","description":"Il existe plusieurs biblioth\u00e8ques open source contenant des mod\u00e8les pr\u00e9-entra\u00een\u00e9s permettant aux Data Scientists et aux entreprises d\u2019utiliser des mod\u00e8les existants afin d'\u00e9conomiser en co\u00fbts de calcul et de gagner du temps. Hugging Face est l\u2019une d\u2019entre elles.","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\/hugging-face-transformers-tout-savoir","og_locale":"fr_FR","og_type":"article","og_title":"Hugging Face Transformers : l\u2019API au service du NLP","og_url":"https:\/\/datascientest.com\/hugging-face-transformers-tout-savoir","og_site_name":"Formation Data Science | DataScientest.com","article_published_time":"2023-06-07T08:18:34+00:00","article_modified_time":"2023-11-09T13:52:12+00:00","og_image":[{"width":1024,"height":563,"url":"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/04\/Illu_BLOG__hugging_face.png","type":"image\/png"}],"author":"Raphael Kassel","twitter_card":"summary_large_image","twitter_misc":{"\u00c9crit par":"Raphael Kassel","Dur\u00e9e de lecture estim\u00e9e":"8 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/datascientest.com\/hugging-face-transformers-tout-savoir","url":"https:\/\/datascientest.com\/hugging-face-transformers-tout-savoir","name":"Hugging Face Transformers : l\u2019API au service du NLP","isPartOf":{"@id":"https:\/\/datascientest.com\/#website"},"datePublished":"2023-06-07T08:18:34+00:00","dateModified":"2023-11-09T13:52:12+00:00","author":{"@id":"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1"},"description":"Il existe plusieurs biblioth\u00e8ques open source contenant des mod\u00e8les pr\u00e9-entra\u00een\u00e9s permettant aux Data Scientists et aux entreprises d\u2019utiliser des mod\u00e8les existants afin d'\u00e9conomiser en co\u00fbts de calcul et de gagner du temps. Hugging Face est l\u2019une d\u2019entre elles.","inLanguage":"fr-FR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/datascientest.com\/hugging-face-transformers-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\/197623"}],"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=197623"}],"version-history":[{"count":0,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/posts\/197623\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/media\/197625"}],"wp:attachment":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/media?parent=197623"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/categories?post=197623"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/tags?post=197623"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}