{"id":226164,"date":"2024-02-20T06:30:00","date_gmt":"2024-02-20T06:30:00","guid":{"rendered":"https:\/\/datascientest.com\/?p=226164"},"modified":"2024-03-14T15:06:40","modified_gmt":"2024-03-14T15:06:40","slug":"fine-tuning-vs-prompt-engineering","status":"publish","type":"post","link":"https:\/\/datascientest.com\/fine-tuning-vs-prompt-engineering","title":{"rendered":"Fine tuning vs Prompt Engineering : Quelles diff\u00e9rences ?"},"content":{"rendered":"\t\t<div data-elementor-type=\"wp-post\" data-elementor-id=\"226164\" class=\"elementor elementor-226164\" data-elementor-post-type=\"post\">\n\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-cd76a62 e-flex e-con-boxed e-con\" data-id=\"cd76a62\" 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-1260f5a elementor-widget elementor-widget-heading\" data-id=\"1260f5a\" 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\">GPT-4, PaLM 2, BARD\u2026 les avanc\u00e9es dans le domaine de l'intelligence artificielle ne cessent de se multiplier. Et surtout, elles deviennent accessibles \u00e0 tous. <\/h2>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-f1e3498 elementor-widget elementor-widget-text-editor\" data-id=\"f1e3498\" 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;\">Plus seulement les <a href=\"https:\/\/datascientest.com\/les-metiers-de-la-data\">experts data<\/a>, mais aussi les chefs d\u2019entreprise, les responsables marketing ou m\u00eame les \u00e9tudiants qui souhaitent un petit coup de pouce. <\/span><\/p><p><span style=\"font-weight: 400;\">Cependant, utiliser ces solutions d\u2019<a href=\"https:\/\/datascientest.com\/comment-lia-generative-pourrait-changer-a-jamais-notre-facon-de-travailler\">IA g\u00e9n\u00e9ratives<\/a> ne s\u2019improvisent pas. \u00c0 travers leur d\u00e9mocratisation, on constate l\u2019importance de savoir optimiser les r\u00e9sultats. C\u2019est l\u00e0 qu\u2019interviennent<\/span><b> le <\/b><a href=\"https:\/\/datascientest.com\/prompt-engineer-tout-savoir\">Prompt Engineering<\/a><b> et le Fine-Tuning.\u00a0 <\/b><span style=\"font-weight: 400;\">Alors de quoi s\u2019agit-il ? Quelles diff\u00e9rences entre ces deux notions ? C\u2019est ce que nous allons d\u00e9couvrir.\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-6c6c793 e-flex e-con-boxed e-con\" data-id=\"6c6c793\" 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-a0129d5 elementor-widget elementor-widget-heading\" data-id=\"a0129d5\" 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\">Qu\u2019est-ce que le Fine-Tuning ?<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-937b6d0 elementor-widget elementor-widget-text-editor\" data-id=\"937b6d0\" 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;\">Malgr\u00e9 leur nom, les <a href=\"https:\/\/datascientest.com\/intelligence-artificielle-definition\">mod\u00e8les d\u2019IA<\/a> ne sont pas \u201cintelligents\u201d. Derri\u00e8re cette suppos\u00e9e intelligence, se cache surtout une capacit\u00e9 \u00e0 trouver des relations dans de vastes et complexes ensembles de donn\u00e9es. Mais les volumes de donn\u00e9es sont tels que les <a href=\"https:\/\/datascientest.com\/large-language-models-tout-savoir\">large language models (LLM)<\/a> d\u00e9pendent des techniques d\u2019optimisation. Parmi lesquels le Fine-Tuning (ou r\u00e9glage fin).\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-453c6ca elementor-widget elementor-widget-heading\" data-id=\"453c6ca\" 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-medium\">D\u00e9finition du Fine-Tuning<\/h4>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-38a1b2c elementor-widget elementor-widget-text-editor\" data-id=\"38a1b2c\" 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;\">Inh\u00e9rente au pr\u00e9-entra\u00eenement des mod\u00e8les, cette technique vise \u00e0 <\/span><b>renforcer la sp\u00e9cialisation d\u2019un mod\u00e8le<\/b><span style=\"font-weight: 400;\"> ayant \u00e9t\u00e9 entra\u00een\u00e9 sur une distribution de donn\u00e9es plus large. Car il existe des mod\u00e8les capables d\u2019effectuer des t\u00e2ches g\u00e9n\u00e9riques. Et d\u2019autres capables de r\u00e9aliser des t\u00e2ches sp\u00e9cifiques. C\u2019est justement dans ce second cas que le Fine-Tuning prend tout son sens. L\u2019id\u00e9e est alors de <\/span><b>r\u00e9pondre sp\u00e9cifiquement aux besoins de l\u2019utilisateur.<\/b><span style=\"font-weight: 400;\"> Par exemple, un chatbot pour le support client d\u2019une entreprise de v\u00eatement ne sera pas entra\u00een\u00e9 de la m\u00eame mani\u00e8re qu\u2019un chatbot charg\u00e9 de g\u00e9n\u00e9rer des pr\u00e9-diagnostics m\u00e9dicaux.\u00a0<\/span><\/p><p><span style=\"font-weight: 400;\">L&rsquo;objectif du Fine-Tuning est alors d<\/span><b>\u2019optimiser les performances d\u2019un mod\u00e8le existant en le r\u00e9-entra\u00eenant <\/b><span style=\"font-weight: 400;\">sur des donn\u00e9es sp\u00e9cifiques. En ajustant les poids et les param\u00e8tres du mod\u00e8le, cette technique permet \u00e0 un syst\u00e8me d\u2019IA de s\u2019adapter \u00e0 des t\u00e2ches plus sp\u00e9cifiques. <\/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-6ad5aaf elementor-widget elementor-widget-heading\" data-id=\"6ad5aaf\" 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-medium\">La r\u00e9ussite du Fine-Tuning<\/h4>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-3d8a870 elementor-widget elementor-widget-text-editor\" data-id=\"3d8a870\" 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;\">Pour am\u00e9liorer les performances du syst\u00e8me d\u2019IA gr\u00e2ce au Fine-Tuning, il convient de pr\u00eater une attention particuli\u00e8re \u00e0 deux param\u00e8tres :\u00a0<\/span><\/p><ul><li style=\"font-weight: 400;\" aria-level=\"1\"><b>La qualit\u00e9 des donn\u00e9es <\/b><span style=\"font-weight: 400;\">: pour r\u00e9pondre \u00e0 des t\u00e2ches sp\u00e9cifiques, les jeux de donn\u00e9es pr\u00e9sent\u00e9s doivent aussi \u00eatre sp\u00e9cifiques.\u00a0<\/span><\/li><li><b>Les \u00e9tapes d\u2019entra\u00eenement : <\/b><span style=\"font-weight: 400;\">il ne s\u2019agit pas seulement d\u2019entra\u00eener le mod\u00e8le \u00e0 contextualiser les donn\u00e9es, mais \u00e0 le guider vers les meilleurs r\u00e9sultats. Pour cela, il est n\u00e9cessaire de mettre en place un syst\u00e8me de retour d\u2019information \u00e0 travers des \u00e9valuations humaines.<\/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-f32fb61 elementor-widget elementor-widget-image\" data-id=\"f32fb61\" 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=\"800\" height=\"575\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/12\/fine-tuning-prompt-engineering1.jpg\" class=\"attachment-large size-large wp-image-226176\" alt=\"\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/12\/fine-tuning-prompt-engineering1.jpg 835w, https:\/\/datascientest.com\/wp-content\/uploads\/2023\/12\/fine-tuning-prompt-engineering1-300x216.jpg 300w, https:\/\/datascientest.com\/wp-content\/uploads\/2023\/12\/fine-tuning-prompt-engineering1-768x552.jpg 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-c3a406f elementor-align-center button-principale elementor-widget elementor-widget-button\" data-id=\"c3a406f\" 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-prompt-engineering\">\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 le Fine-Tuning<\/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-6cfa6ea e-flex e-con-boxed e-con\" data-id=\"6cfa6ea\" 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-ac93778 elementor-widget elementor-widget-heading\" data-id=\"ac93778\" 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\">Qu\u2019est-ce que le Prompt Engineering ?<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-e9b1bf8 elementor-widget elementor-widget-text-editor\" data-id=\"e9b1bf8\" 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;\">Tout comme le Fine-Tuning, le Prompt Engineering vise aussi \u00e0 am\u00e9liorer les <a href=\"https:\/\/datascientest.com\/machine-learning-tout-savoir\">performances du mod\u00e8le de machine learning<\/a>. Mais ici, on ne se concentre pas tant sur l\u2019entra\u00eenement du mod\u00e8le, mais plut\u00f4t sur <\/span><b>les r\u00e9sultats fournis gr\u00e2ce aux prompts.\u00a0<\/b><\/p><p><span style=\"font-weight: 400;\">Pour rappel, les prompts (\u00e9galement appel\u00e9s invites ou messages guid\u00e9s en fran\u00e7ais) correspondent aux demandes d\u2019un utilisateur \u00e0 une application d\u2019intelligence artificielle g\u00e9n\u00e9rative (<a href=\"https:\/\/datascientest.com\/gpt-3-revolutionnaire-ou-surmediatise\">comme ChatGPT<\/a>). Et c\u2019est justement la qualit\u00e9 de ces demandes (les inputs prompts) qui d\u00e9termine la qualit\u00e9 du r\u00e9sultat (l\u2019output).\u00a0<\/span><\/p><p><span style=\"font-weight: 400;\">Concr\u00e8tement, lorsque vous saisissez une requ\u00eate, l\u2019intelligence artificielle d\u00e9compose le prompt pour en comprendre le sens et l\u2019intention. Si vous posez une question assez g\u00e9n\u00e9rale, le syst\u00e8me d\u2019IA vous fournira une r\u00e9ponse en fonction des donn\u00e9es qu\u2019il a \u00e0 sa disposition. Mais comme il n\u2019a pas forc\u00e9ment bien saisi votre intention, il pourra r\u00e9pondre \u00e0 c\u00f4t\u00e9. \u00c0 l\u2019inverse, si vous fournissez un prompt pr\u00e9cis en expliquant votre intention et le contexte, l\u2019intelligence artificielle sera plus en mesure de r\u00e9pondre \u00e0 votre demande.\u00a0<\/span><\/p><p><span style=\"font-weight: 400;\">Ainsi, le Prompt Engineering consiste <\/span><b>\u00e0 guider les r\u00e9ponses des mod\u00e8les de ML en formulant des requ\u00eates sp\u00e9cifiques. <\/b><span style=\"font-weight: 400;\">Pour r\u00e9ussir vos prompts, n\u2019h\u00e9sitez pas \u00e0 poser des questions de plus en plus pointues et \u00e0 tester diff\u00e9rentes fa\u00e7ons de formuler les instructions. <\/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-212e917 e-flex e-con-boxed e-con\" data-id=\"212e917\" 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-e179c81 elementor-widget elementor-widget-heading\" data-id=\"e179c81\" 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\">Prompt Engineering vs Fine-Tuning - Quelles diff\u00e9rences ?<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-ca4a673 elementor-widget elementor-widget-text-editor\" data-id=\"ca4a673\" 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;\">Fine-Tuning et Prompt Engineering sont deux techniques d\u2019optimisation de l\u2019IA. Cependant, il convient de conna\u00eetre les diff\u00e9rences entre ces deux m\u00e9thodes :\u00a0<\/span><\/p><ul><li style=\"font-weight: 400;\" aria-level=\"1\"><b>L\u2019objectif <\/b><span style=\"font-weight: 400;\">: le Prompt Engineering se concentre davantage sur la production de r\u00e9sultats pertinents, alors que le Fine-Tuning vise \u00e0 am\u00e9liorer les performances du mod\u00e8le de machine learning pour accomplir certaines t\u00e2ches.<\/span><\/li><li style=\"font-weight: 400;\" aria-level=\"1\"><b>La m\u00e9thode : <\/b><span style=\"font-weight: 400;\">le Prompt Engineering est bas\u00e9 sur les indications en entr\u00e9e. Celles-ci doivent \u00eatre plus d\u00e9taill\u00e9es pour obtenir de meilleurs r\u00e9sultats. De son c\u00f4t\u00e9, la performance du Fine-Tuning repose sur l\u2019entra\u00eenement d\u2019un mod\u00e8le et l\u2019int\u00e9gration de nouvelles donn\u00e9es sp\u00e9cifiques.\u00a0\u00a0<\/span><\/li><li style=\"font-weight: 400;\" aria-level=\"1\"><b>Le contr\u00f4le <\/b><span style=\"font-weight: 400;\">: avec le Prompt Engineering, c\u2019est l\u2019utilisateur qui garde pleinement le contr\u00f4le des r\u00e9sultats. \u00c0 l\u2019inverse, le Fine-Tuning vise \u00e0 donner davantage d\u2019autonomie \u00e0 l\u2019ordinateur pour qu\u2019il soit capable de produire les r\u00e9sultats souhait\u00e9s.\u00a0<\/span><\/li><li><b>Les moyens <\/b><span style=\"font-weight: 400;\">: le Prompt Engineering ne n\u00e9cessite pas forc\u00e9ment de ressources. D\u2019autant plus que de nombreuses applications d\u2019IA g\u00e9n\u00e9rative sont gratuites. Ainsi, toute personne peut affiner ses prompts pour optimiser ses r\u00e9sultats. Au contraire, le Fine-Tuning n\u00e9cessite des ressources importantes, \u00e0 la fois pour l\u2019entra\u00eenement des mod\u00e8les, mais aussi pour l\u2019ajout de grands volumes de donn\u00e9es. <\/span><\/li><\/ul>\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-00ceac5 e-flex e-con-boxed e-con\" data-id=\"00ceac5\" 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-201329b elementor-widget elementor-widget-heading\" data-id=\"201329b\" 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\">Comment ma\u00eetriser le Fine-Tuning et le Prompt Engineering ?<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-29c6166 elementor-widget elementor-widget-text-editor\" data-id=\"29c6166\" 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;\">Bien que distincts, le Fine-Tuning et le Prompt Engineering partagent un objectif commun : celui d&rsquo;am\u00e9liorer la pr\u00e9cision et la pertinence des r\u00e9ponses g\u00e9n\u00e9r\u00e9es par les mod\u00e8les d\u2019IA g\u00e9n\u00e9rative. Pour les experts data, il est plus que n\u00e9cessaire de ma\u00eetriser ces techniques. C\u2019est l\u00e0 qu\u2019intervient DataScientest. \u00c0 travers notre programme de formation, vous d\u00e9couvrirez les meilleures techniques pour affiner les r\u00e9sultats des mod\u00e8les de langage. <a href=\"https:\/\/datascientest.com\/formation-data-scientist\"><span style=\"font-weight: 400;\">Rejoignez-nous !<\/span><\/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-1bbea85 elementor-widget elementor-widget-image\" data-id=\"1bbea85\" 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=\"644\" height=\"600\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/12\/fine-tuning-prompt-engineering2.jpg\" class=\"attachment-large size-large wp-image-226180\" alt=\"\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/12\/fine-tuning-prompt-engineering2.jpg 644w, https:\/\/datascientest.com\/wp-content\/uploads\/2023\/12\/fine-tuning-prompt-engineering2-300x280.jpg 300w\" sizes=\"(max-width: 644px) 100vw, 644px\" \/>\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-2580324 elementor-align-center button-principale elementor-widget elementor-widget-button\" data-id=\"2580324\" 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-prompt-engineering\">\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\">Commencer une formation en Prompt Engineering<\/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<div class=\"elementor-element elementor-element-c884109 elementor-widget elementor-widget-html\" data-id=\"c884109\" 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\u2019est-ce que le Fine-Tuning ?\",\r\n    \"acceptedAnswer\": {\r\n      \"@type\": \"Answer\",\r\n      \"text\": \"Malgr\u00e9 leur nom, les mod\u00e8les d\u2019IA ne sont pas \u201cintelligents\u201d. Derri\u00e8re cette suppos\u00e9e intelligence, se cache surtout une capacit\u00e9 \u00e0 trouver des relations dans de vastes et complexes ensembles de donn\u00e9es. Mais les volumes de donn\u00e9es sont tels que les large language models (LLM) d\u00e9pendent des techniques d\u2019optimisation. Parmi lesquels le Fine-Tuning (ou r\u00e9glage fin).\"\r\n    }\r\n  },{\r\n    \"@type\": \"Question\",\r\n    \"name\": \"Qu\u2019est-ce que le Prompt Engineering ?\",\r\n    \"acceptedAnswer\": {\r\n      \"@type\": \"Answer\",\r\n      \"text\": \"Tout comme le Fine-Tuning, le Prompt Engineering vise aussi \u00e0 am\u00e9liorer les performances du mod\u00e8le de machine learning. Mais ici, on ne se concentre pas tant sur l\u2019entra\u00eenement du mod\u00e8le, mais plut\u00f4t sur les r\u00e9sultats fournis gr\u00e2ce aux prompts.\"\r\n    }\r\n  },{\r\n    \"@type\": \"Question\",\r\n    \"name\": \"Prompt Engineering vs Fine-Tuning - Quelles diff\u00e9rences ?\",\r\n    \"acceptedAnswer\": {\r\n      \"@type\": \"Answer\",\r\n      \"text\": \"Fine-Tuning et Prompt Engineering sont deux techniques d\u2019optimisation de l\u2019IA. Cependant, il convient de conna\u00eetre les diff\u00e9rences entre ces deux m\u00e9thodes :\r\n      L\u2019objectif : le Prompt Engineering se concentre davantage sur la production de r\u00e9sultats pertinents, alors que le Fine-Tuning vise \u00e0 am\u00e9liorer les performances du mod\u00e8le de machine learning pour accomplir certaines t\u00e2ches.\r\n      La m\u00e9thode : le Prompt Engineering est bas\u00e9 sur les indications en entr\u00e9e. Celles-ci doivent \u00eatre plus d\u00e9taill\u00e9es pour obtenir de meilleurs r\u00e9sultats. De son c\u00f4t\u00e9, la performance du Fine-Tuning repose sur l\u2019entra\u00eenement d\u2019un mod\u00e8le et l\u2019int\u00e9gration de nouvelles donn\u00e9es sp\u00e9cifiques.\r\n      Le contr\u00f4le : avec le Prompt Engineering, c\u2019est l\u2019utilisateur qui garde pleinement le contr\u00f4le des r\u00e9sultats. \u00c0 l\u2019inverse, le Fine-Tuning vise \u00e0 donner davantage d\u2019autonomie \u00e0 l\u2019ordinateur pour qu\u2019il soit capable de produire les r\u00e9sultats souhait\u00e9s.\r\n      Les moyens : le Prompt Engineering ne n\u00e9cessite pas forc\u00e9ment de ressources. D\u2019autant plus que de nombreuses applications d\u2019IA g\u00e9n\u00e9rative sont gratuites. Ainsi, toute personne peut affiner ses prompts pour optimiser ses r\u00e9sultats. Au contraire, le Fine-Tuning n\u00e9cessite des ressources importantes, \u00e0 la fois pour l\u2019entra\u00eenement des mod\u00e8les, mais aussi pour l\u2019ajout de grands volumes de donn\u00e9es.\"\r\n    }\r\n  }]\r\n}\r\n<\/script>\r\n\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>GPT-4, PaLM 2, BARD\u2026 les avanc\u00e9es dans le domaine de l&rsquo;intelligence artificielle ne cessent de se multiplier. Et surtout, elles deviennent accessibles \u00e0 tous. Plus seulement les experts data, mais aussi les chefs d\u2019entreprise, les responsables marketing ou m\u00eame les \u00e9tudiants qui souhaitent un petit coup de pouce. Cependant, utiliser ces solutions d\u2019IA g\u00e9n\u00e9ratives ne [&hellip;]<\/p>\n","protected":false},"author":77,"featured_media":226166,"comment_status":"open","ping_status":"open","sticky":false,"template":"elementor_theme","format":"standard","meta":{"content-type":"","inline_featured_image":false},"categories":[2500],"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>Fine tuning vs Prompt Engineering : Quelles diff\u00e9rences ?<\/title>\n<meta name=\"description\" content=\"Les performances de IA d\u00e9pendent en grande partie des m\u00e9thodes d\u2019optimisation qui affinent ses r\u00e9sultats, notamment le fine tuning et le prompt engineering.\" \/>\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\/fine-tuning-vs-prompt-engineering\" \/>\n<meta property=\"og:locale\" content=\"fr_FR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Fine tuning vs Prompt Engineering : Quelles diff\u00e9rences ?\" \/>\n<meta property=\"og:url\" content=\"https:\/\/datascientest.com\/fine-tuning-vs-prompt-engineering\" \/>\n<meta property=\"og:site_name\" content=\"Formation Data Science | DataScientest.com\" \/>\n<meta property=\"article:published_time\" content=\"2024-02-20T06:30:00+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2024-03-14T15:06:40+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/12\/Fine-Tuning-vs-Prompt-Engineering.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"1050\" \/>\n\t<meta property=\"og:image:height\" content=\"600\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\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=\"6 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/datascientest.com\/fine-tuning-vs-prompt-engineering\",\"url\":\"https:\/\/datascientest.com\/fine-tuning-vs-prompt-engineering\",\"name\":\"Fine tuning vs Prompt Engineering : Quelles diff\u00e9rences ?\",\"isPartOf\":{\"@id\":\"https:\/\/datascientest.com\/#website\"},\"datePublished\":\"2024-02-20T06:30:00+00:00\",\"dateModified\":\"2024-03-14T15:06:40+00:00\",\"author\":{\"@id\":\"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1\"},\"description\":\"Les performances de IA d\u00e9pendent en grande partie des m\u00e9thodes d\u2019optimisation qui affinent ses r\u00e9sultats, notamment le fine tuning et le prompt engineering.\",\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/datascientest.com\/fine-tuning-vs-prompt-engineering\"]}]},{\"@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":"Fine tuning vs Prompt Engineering : Quelles diff\u00e9rences ?","description":"Les performances de IA d\u00e9pendent en grande partie des m\u00e9thodes d\u2019optimisation qui affinent ses r\u00e9sultats, notamment le fine tuning et le prompt engineering.","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\/fine-tuning-vs-prompt-engineering","og_locale":"fr_FR","og_type":"article","og_title":"Fine tuning vs Prompt Engineering : Quelles diff\u00e9rences ?","og_url":"https:\/\/datascientest.com\/fine-tuning-vs-prompt-engineering","og_site_name":"Formation Data Science | DataScientest.com","article_published_time":"2024-02-20T06:30:00+00:00","article_modified_time":"2024-03-14T15:06:40+00:00","og_image":[{"width":1050,"height":600,"url":"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/12\/Fine-Tuning-vs-Prompt-Engineering.jpg","type":"image\/jpeg"}],"author":"Raphael Kassel","twitter_card":"summary_large_image","twitter_misc":{"\u00c9crit par":"Raphael Kassel","Dur\u00e9e de lecture estim\u00e9e":"6 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/datascientest.com\/fine-tuning-vs-prompt-engineering","url":"https:\/\/datascientest.com\/fine-tuning-vs-prompt-engineering","name":"Fine tuning vs Prompt Engineering : Quelles diff\u00e9rences ?","isPartOf":{"@id":"https:\/\/datascientest.com\/#website"},"datePublished":"2024-02-20T06:30:00+00:00","dateModified":"2024-03-14T15:06:40+00:00","author":{"@id":"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1"},"description":"Les performances de IA d\u00e9pendent en grande partie des m\u00e9thodes d\u2019optimisation qui affinent ses r\u00e9sultats, notamment le fine tuning et le prompt engineering.","inLanguage":"fr-FR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/datascientest.com\/fine-tuning-vs-prompt-engineering"]}]},{"@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\/226164"}],"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=226164"}],"version-history":[{"count":0,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/posts\/226164\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/media\/226166"}],"wp:attachment":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/media?parent=226164"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/categories?post=226164"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/tags?post=226164"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}