{"id":2314,"date":"2026-05-11T11:21:38","date_gmt":"2026-05-11T09:21:38","guid":{"rendered":"https:\/\/askem.eu\/?p=2314"},"modified":"2026-05-11T11:21:44","modified_gmt":"2026-05-11T09:21:44","slug":"colpali-se-passer-docr-pour-la-recherche","status":"publish","type":"post","link":"https:\/\/askem.eu\/en\/2026\/05\/11\/colpali-se-passer-docr-pour-la-recherche\/","title":{"rendered":"ColPali, se passer d&rsquo;OCR pour la recherche"},"content":{"rendered":"<h2 class=\"wp-block-heading\">ColPali&nbsp;: la recherche documentaire visuelle qui rend l&rsquo;OCR obsol\u00e8te&nbsp;?<\/h2>\n\n\n\n<p>La majorit\u00e9 des pipelines RAG en production reposent sur un pr\u00e9suppos\u00e9 fragile&nbsp;: pour exploiter un PDF, il faut d&rsquo;abord le transformer en texte. On encha\u00eene donc OCR, segmentation, nettoyage, extraction de tableaux, parfois une \u00e9tape de mise en page. Chaque maillon ajoute du bruit, perd de l&rsquo;information visuelle (colonnes, encadr\u00e9s, sch\u00e9mas, signatures, logos), et n\u00e9cessite une maintenance permanente. <strong><a href=\"https:\/\/github.com\/illuin-tech\/colpali\">ColPali<\/a><\/strong>, publi\u00e9 en 2024 par l&rsquo;\u00e9quipe d&rsquo;Illuin Technology et l&rsquo;Universit\u00e9 Paris Sciences et Lettres, propose une autre voie&nbsp;: indexer directement l&rsquo;image de la page, sans OCR, et laisser un mod\u00e8le de vision embarquer ce que l&rsquo;\u0153il voit. Le mod\u00e8le est sous licence MIT, les poids sont publi\u00e9s sur Hugging Face, et les benchmarks sont publics.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Le verrou que ColPali fait sauter<\/h3>\n\n\n\n<p>Sur des documents bureautiques classiques, un pipeline OCR + chunking texte fonctionne raisonnablement. Sur des documents r\u00e9els d&rsquo;organisation publique, de cabinet d&rsquo;\u00e9tudes, de bureau d&rsquo;ing\u00e9nierie, l&rsquo;OCR atteint vite ses limites&nbsp;: tableaux financiers, formulaires complexes, plans, captures d&rsquo;\u00e9cran, sch\u00e9mas l\u00e9gend\u00e9s, factures scann\u00e9es de travers, anciens rapports en colonnes, notes manuscrites, infographies. Le contenu visuel porte une part de l&rsquo;information qu&rsquo;aucun extracteur texte ne r\u00e9cup\u00e8re proprement. ColPali contourne le probl\u00e8me en court-circuitant l&rsquo;\u00e9tape OCR&nbsp;: la page est trait\u00e9e comme une image, d\u00e9coup\u00e9e en patches par un encodeur visuel, et chaque patch re\u00e7oit un embedding qui capture \u00e0 la fois le texte rendu et la structure visuelle.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Comment \u00e7a marche, sans entrer dans la cuisine<\/h3>\n\n\n\n<p>ColPali combine deux id\u00e9es d\u00e9j\u00e0 \u00e9prouv\u00e9es s\u00e9par\u00e9ment, ColBERT c\u00f4t\u00e9 retrieval et PaliGemma c\u00f4t\u00e9 vision, et les fusionne. Au lieu de produire un seul vecteur par page, le mod\u00e8le produit une grille de vecteurs, un par patch d&rsquo;image. Au moment de la recherche, la requ\u00eate textuelle est elle aussi tokenis\u00e9e en plusieurs vecteurs, et le score d&rsquo;une page est la somme des meilleures correspondances entre chaque token de la requ\u00eate et chaque patch de la page. C&rsquo;est l&rsquo;approche dite <em>late interaction<\/em> de ColBERT, transpos\u00e9e \u00e0 la vision. Pratiquement, cela permet \u00e0 une question sur un montant d&rsquo;aller chercher exactement la cellule du tableau o\u00f9 ce montant figure, et \u00e0 une question sur un sch\u00e9ma de pointer la zone illustr\u00e9e correspondante.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Mise en place pratique<\/h3>\n\n\n\n<p>La biblioth\u00e8que de r\u00e9f\u00e9rence est <code>colpali-engine<\/code>, distribu\u00e9e sur PyPI sous licence MIT. L&rsquo;installation et la premi\u00e8re indexation tiennent en quelques lignes&nbsp;:<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>from colpali_engine.models import ColPali, ColPaliProcessor\nfrom pdf2image import convert_from_path\n\nmodele = ColPali.from_pretrained(\"vidore\/colpali-v1.3\")\nprocessor = ColPaliProcessor.from_pretrained(\"vidore\/colpali-v1.3\")\n\npages = convert_from_path(\"rapport.pdf\", dpi=150)\nembeddings_pages = modele(**processor.process_images(pages))\n\nrequete = processor.process_queries(&#91;\"quel est le montant total de la subvention&nbsp;?\"])\nembeddings_req = modele(**requete)\n\nscores = processor.score_multi_vector(embeddings_req, embeddings_pages)<\/code><\/pre>\n\n\n\n<p>Aucune passe OCR, aucun parseur de tableau, aucune r\u00e8gle m\u00e9tier sur la mise en page. Le pipeline d&rsquo;ingestion se r\u00e9sume \u00e0&nbsp;: convertir le PDF en images de page, calculer les embeddings, les stocker. C\u00f4t\u00e9 infrastructure, une carte avec 16 Go de VRAM suffit largement pour indexer plusieurs milliers de pages par heure&nbsp;; pour servir en lecture, le surco\u00fbt m\u00e9moire reste raisonnable gr\u00e2ce \u00e0 la quantification 8 bits propos\u00e9e nativement.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Stockage et passage \u00e0 l&rsquo;\u00e9chelle<\/h3>\n\n\n\n<p>La grille de vecteurs par page consomme plus que les approches mono-vecteur classiques&nbsp;: compter environ 1 000 vecteurs de dimension 128 par page, soit quelques centaines de kilo-octets par page apr\u00e8s quantification. Pour un corpus jusqu&rsquo;\u00e0 100 000 pages, un stockage local en parquet ou un index d\u00e9di\u00e9 comme <strong>Vespa<\/strong>, <strong>Qdrant<\/strong> en mode multi-vecteur, ou <strong>plaid-x<\/strong> tient confortablement sur une machine de production. Au-del\u00e0, on bascule sur un index distribu\u00e9 ou sur la variante <strong>ColPali tensor compression<\/strong> qui r\u00e9duit l&#8217;empreinte d&rsquo;un facteur cinq sans d\u00e9gradation significative des scores.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Le benchmark ViDoRe et ce qu&rsquo;il dit vraiment<\/h3>\n\n\n\n<p>L&rsquo;\u00e9quipe de ColPali a publi\u00e9 en parall\u00e8le <strong><a href=\"https:\/\/github.com\/illuin-tech\/vidore-benchmark\">ViDoRe<\/a><\/strong>, un benchmark d&rsquo;\u00e9valuation de retrieval documentaire visuel, couvrant rapports financiers, articles scientifiques, infographies, documents administratifs en plusieurs langues dont le fran\u00e7ais. Les r\u00e9sultats publics placent ColPali devant les pipelines OCR + BM25 et OCR + embeddings sur la quasi-totalit\u00e9 des sous-t\u00e2ches, avec un \u00e9cart particuli\u00e8rement marqu\u00e9 sur les documents riches en tableaux et en sch\u00e9mas. L&rsquo;\u00e9cart se r\u00e9duit sur les documents purement textuels o\u00f9 l&rsquo;OCR fait correctement son travail, ce qui sugg\u00e8re une strat\u00e9gie hybride&nbsp;: ColPali sur les documents visuellement complexes, embeddings texte classiques ailleurs, avec un routeur de requ\u00eate en amont.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Ce que \u00e7a change pour un projet RAG c\u00f4t\u00e9 Askem<\/h3>\n\n\n\n<p>Trois b\u00e9n\u00e9fices se concr\u00e9tisent rapidement sur les projets que l&rsquo;on rencontre. D&rsquo;abord, le co\u00fbt d&rsquo;ingestion s&rsquo;effondre&nbsp;: plus de cha\u00eene fragile OCR + extraction de tableaux + nettoyage, donc moins d&rsquo;incidents et un temps d&rsquo;int\u00e9gration de nouveaux corpus divis\u00e9 par cinq. Ensuite, la qualit\u00e9 de r\u00e9ponse s&rsquo;am\u00e9liore sur les documents que l&rsquo;OCR maltraitait&nbsp;: rapports d&rsquo;\u00e9tudes, plans, formulaires, sch\u00e9mas, o\u00f9 ColPali r\u00e9cup\u00e8re le contexte visuel que le texte seul n&rsquo;expose pas. Enfin, la tra\u00e7abilit\u00e9 progresse&nbsp;: on peut surligner la zone exacte de la page utilis\u00e9e pour la r\u00e9ponse, ce que beaucoup de r\u00e9f\u00e9rents m\u00e9tier r\u00e9clament dans les SI publics ou r\u00e9glement\u00e9s.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Limites \u00e0 conna\u00eetre avant de se lancer<\/h3>\n\n\n\n<p>ColPali ne dispense pas d&rsquo;\u00e9valuer son pipeline avec un outil comme <a href=\"https:\/\/github.com\/vibrantlabsai\/ragas\">RAGAS<\/a>. Le mod\u00e8le h\u00e9rite des biais de PaliGemma sur les langues moins repr\u00e9sent\u00e9es, et la qualit\u00e9 reste meilleure en anglais qu&rsquo;en fran\u00e7ais, m\u00eame si l&rsquo;\u00e9cart se r\u00e9duit avec les versions successives. La taille de l&rsquo;index multi-vecteur est un poste de co\u00fbt \u00e0 int\u00e9grer d\u00e8s la conception. Enfin, ColPali ne remplace pas les briques aval&nbsp;: il faut toujours un LLM pour g\u00e9n\u00e9rer la r\u00e9ponse \u00e0 partir des pages remont\u00e9es, id\u00e9alement multimodal pour pouvoir lire l&rsquo;image directement, comme Claude ou un mod\u00e8le ouvert servi via vLLM.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Le lien avec Docling&nbsp;: compl\u00e9ment, pas concurrent<\/h3>\n\n\n\n<p>ColPali ne remplace pas <strong>Docling<\/strong>, pr\u00e9sent\u00e9 ici dans un article pr\u00e9c\u00e9dent. Les deux outils r\u00e8glent des probl\u00e8mes diff\u00e9rents et se combinent tr\u00e8s bien. Docling est un <em>extracteur<\/em>&nbsp;: il transforme un PDF en repr\u00e9sentation textuelle structur\u00e9e (markdown hi\u00e9rarchis\u00e9, tableaux, m\u00e9tadonn\u00e9es), id\u00e9ale pour b\u00e2tir un index texte classique, alimenter un moteur de recherche plein-texte, ou nourrir un LLM avec des passages propres. ColPali est un <em>moteur de retrieval visuel<\/em>&nbsp;: il court-circuite l&rsquo;extraction et indexe directement l&rsquo;image de la page, ce qui pr\u00e9serve la mise en page, les sch\u00e9mas et la structure visuelle des tableaux. <\/p>\n\n\n\n<p>Dans un projet r\u00e9el, la combinaison la plus robuste consiste \u00e0 faire tourner <a href=\"https:\/\/askem.eu\/en\/2026\/04\/10\/docling-convertir-pdf-docx-et-images-en-donnees-structurees-pour-ses-pipelines-rag\/\" type=\"post\" id=\"2216\">Docling<\/a> sur le corpus textuel courant, ColPali sur le sous-ensemble visuellement complexe (rapports d&rsquo;\u00e9tudes, formulaires, plans, infographies), et \u00e0 fusionner les scores via un retrieval hybride. <\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Ce qu&rsquo;il faut retenir<\/h3>\n\n\n\n<p>ColPali n&rsquo;est pas un gadget de recherche, c&rsquo;est un changement d&rsquo;architecture pour le retrieval documentaire. Pour tout corpus o\u00f9 la mise en page porte du sens, c&rsquo;est aujourd&rsquo;hui la voie open source la plus solide pour b\u00e2tir un RAG fiable, tra\u00e7able, et nettement moins fragile que les pipelines OCR historiques. \u00c0 tester d\u00e8s la phase POC, en parall\u00e8le d&rsquo;une approche texte classique, sur un \u00e9chantillon repr\u00e9sentatif de documents r\u00e9els.<\/p>\n\n\n\n<p><\/p>","protected":false},"excerpt":{"rendered":"<p>ColPali&nbsp;: la recherche documentaire visuelle qui rend l&rsquo;OCR obsol\u00e8te&nbsp;? La majorit\u00e9 des pipelines RAG en production reposent sur un pr\u00e9suppos\u00e9 fragile&nbsp;: pour exploiter un PDF, il faut d&rsquo;abord le transformer en texte. On encha\u00eene donc OCR, segmentation, nettoyage, extraction de tableaux, parfois une \u00e9tape de mise en page. Chaque maillon ajoute du bruit, perd de [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":2315,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"ocean_post_layout":"","ocean_both_sidebars_style":"","ocean_both_sidebars_content_width":0,"ocean_both_sidebars_sidebars_width":0,"ocean_sidebar":"","ocean_second_sidebar":"","ocean_disable_margins":"enable","ocean_add_body_class":"","ocean_shortcode_before_top_bar":"","ocean_shortcode_after_top_bar":"","ocean_shortcode_before_header":"","ocean_shortcode_after_header":"","ocean_has_shortcode":"","ocean_shortcode_after_title":"","ocean_shortcode_before_footer_widgets":"","ocean_shortcode_after_footer_widgets":"","ocean_shortcode_before_footer_bottom":"","ocean_shortcode_after_footer_bottom":"","ocean_display_top_bar":"default","ocean_display_header":"default","ocean_header_style":"","ocean_center_header_left_menu":"","ocean_custom_header_template":"","ocean_custom_logo":0,"ocean_custom_retina_logo":0,"ocean_custom_logo_max_width":0,"ocean_custom_logo_tablet_max_width":0,"ocean_custom_logo_mobile_max_width":0,"ocean_custom_logo_max_height":0,"ocean_custom_logo_tablet_max_height":0,"ocean_custom_logo_mobile_max_height":0,"ocean_header_custom_menu":"","ocean_menu_typo_font_family":"","ocean_menu_typo_font_subset":"","ocean_menu_typo_font_size":0,"ocean_menu_typo_font_size_tablet":0,"ocean_menu_typo_font_size_mobile":0,"ocean_menu_typo_font_size_unit":"px","ocean_menu_typo_font_weight":"","ocean_menu_typo_font_weight_tablet":"","ocean_menu_typo_font_weight_mobile":"","ocean_menu_typo_transform":"","ocean_menu_typo_transform_tablet":"","ocean_menu_typo_transform_mobile":"","ocean_menu_typo_line_height":0,"ocean_menu_typo_line_height_tablet":0,"ocean_menu_typo_line_height_mobile":0,"ocean_menu_typo_line_height_unit":"","ocean_menu_typo_spacing":0,"ocean_menu_typo_spacing_tablet":0,"ocean_menu_typo_spacing_mobile":0,"ocean_menu_typo_spacing_unit":"","ocean_menu_link_color":"","ocean_menu_link_color_hover":"","ocean_menu_link_color_active":"","ocean_menu_link_background":"","ocean_menu_link_hover_background":"","ocean_menu_link_active_background":"","ocean_menu_social_links_bg":"","ocean_menu_social_hover_links_bg":"","ocean_menu_social_links_color":"","ocean_menu_social_hover_links_color":"","ocean_disable_title":"default","ocean_disable_heading":"default","ocean_post_title":"","ocean_post_subheading":"","ocean_post_title_style":"","ocean_post_title_background_color":"","ocean_post_title_background":0,"ocean_post_title_bg_image_position":"","ocean_post_title_bg_image_attachment":"","ocean_post_title_bg_image_repeat":"","ocean_post_title_bg_image_size":"","ocean_post_title_height":0,"ocean_post_title_bg_overlay":0.5,"ocean_post_title_bg_overlay_color":"","ocean_disable_breadcrumbs":"default","ocean_breadcrumbs_color":"","ocean_breadcrumbs_separator_color":"","ocean_breadcrumbs_links_color":"","ocean_breadcrumbs_links_hover_color":"","ocean_display_footer_widgets":"default","ocean_display_footer_bottom":"default","ocean_custom_footer_template":"","osh_disable_topbar_sticky":"default","osh_disable_header_sticky":"default","osh_sticky_header_style":"default","osh_sticky_header_effect":"","osh_custom_sticky_logo":0,"osh_custom_retina_sticky_logo":0,"osh_custom_sticky_logo_height":0,"osh_background_color":"","osh_links_color":"","osh_links_hover_color":"","osh_links_active_color":"","osh_links_bg_color":"","osh_links_hover_bg_color":"","osh_links_active_bg_color":"","osh_menu_social_links_color":"","osh_menu_social_hover_links_color":"","ocean_post_oembed":"","ocean_post_self_hosted_media":"","ocean_post_video_embed":"","ocean_link_format":"","ocean_link_format_target":"self","ocean_quote_format":"","ocean_quote_format_link":"post","ocean_gallery_link_images":"on","ocean_gallery_id":[],"footnotes":""},"categories":[16],"tags":[],"class_list":["post-2314","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai","entry","has-media"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v27.6 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>ColPali, se passer d&#039;OCR pour la recherche - askem<\/title>\n<meta name=\"description\" content=\"ASKEM BUREAU D&#039;\u00c9TUDES ET DE FORMATION NUM\u00c9RIQUE. Nous vous assistons dans la transformation num\u00e9rique de vos outils, services et organisations tout en pla\u00e7ant l\u2019humain au c\u0153ur de notre d\u00e9marche d\u2019accompagnement.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/askem.eu\/en\/2026\/05\/11\/colpali-se-passer-docr-pour-la-recherche\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"ColPali, se passer d&#039;OCR pour la recherche - askem\" \/>\n<meta property=\"og:description\" content=\"ASKEM BUREAU D&#039;\u00c9TUDES ET DE FORMATION NUM\u00c9RIQUE. Nous vous assistons dans la transformation num\u00e9rique de vos outils, services et organisations tout en pla\u00e7ant l\u2019humain au c\u0153ur de notre d\u00e9marche d\u2019accompagnement.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/askem.eu\/en\/2026\/05\/11\/colpali-se-passer-docr-pour-la-recherche\/\" \/>\n<meta property=\"og:site_name\" content=\"askem\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/fb.me\/askem.eu\" \/>\n<meta property=\"article:published_time\" content=\"2026-05-11T09:21:38+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2026-05-11T09:21:44+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/mlpi0fxo3sth.i.optimole.com\/cb:3obA.c61\/w:auto\/h:auto\/q:mauto\/f:best\/https:\/\/askem.eu\/wp-content\/uploads\/2026\/05\/sujet-askem-2026-05-08.png\" \/>\n\t<meta property=\"og:image:width\" content=\"1600\" \/>\n\t<meta property=\"og:image:height\" content=\"1200\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"author\" content=\"askemadmin\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"askemadmin\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\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\":\"Article\",\"@id\":\"https:\\\/\\\/askem.eu\\\/2026\\\/05\\\/11\\\/colpali-se-passer-docr-pour-la-recherche\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/askem.eu\\\/2026\\\/05\\\/11\\\/colpali-se-passer-docr-pour-la-recherche\\\/\"},\"author\":{\"name\":\"askemadmin\",\"@id\":\"https:\\\/\\\/askem.eu\\\/#\\\/schema\\\/person\\\/8bbee74ab9a977d56bf4826662e9d2e9\"},\"headline\":\"ColPali, se passer d&rsquo;OCR pour la recherche\",\"datePublished\":\"2026-05-11T09:21:38+00:00\",\"dateModified\":\"2026-05-11T09:21:44+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/askem.eu\\\/2026\\\/05\\\/11\\\/colpali-se-passer-docr-pour-la-recherche\\\/\"},\"wordCount\":1238,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\\\/\\\/askem.eu\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/askem.eu\\\/2026\\\/05\\\/11\\\/colpali-se-passer-docr-pour-la-recherche\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\/\\/askem.eu\\/wp-content\\/uploads\\/2026\\/05\\/sujet-askem-2026-05-08.png\",\"articleSection\":[\"AI\"],\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/askem.eu\\\/2026\\\/05\\\/11\\\/colpali-se-passer-docr-pour-la-recherche\\\/#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/askem.eu\\\/2026\\\/05\\\/11\\\/colpali-se-passer-docr-pour-la-recherche\\\/\",\"url\":\"https:\\\/\\\/askem.eu\\\/2026\\\/05\\\/11\\\/colpali-se-passer-docr-pour-la-recherche\\\/\",\"name\":\"ColPali, se passer d'OCR pour la recherche - askem\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/askem.eu\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/askem.eu\\\/2026\\\/05\\\/11\\\/colpali-se-passer-docr-pour-la-recherche\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/askem.eu\\\/2026\\\/05\\\/11\\\/colpali-se-passer-docr-pour-la-recherche\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\/\\/askem.eu\\/wp-content\\/uploads\\/2026\\/05\\/sujet-askem-2026-05-08.png\",\"datePublished\":\"2026-05-11T09:21:38+00:00\",\"dateModified\":\"2026-05-11T09:21:44+00:00\",\"description\":\"ASKEM BUREAU D'\u00c9TUDES ET DE FORMATION NUM\u00c9RIQUE. Nous vous assistons dans la transformation num\u00e9rique de vos outils, services et organisations tout en pla\u00e7ant l\u2019humain au c\u0153ur de notre d\u00e9marche d\u2019accompagnement.\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/askem.eu\\\/2026\\\/05\\\/11\\\/colpali-se-passer-docr-pour-la-recherche\\\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/askem.eu\\\/2026\\\/05\\\/11\\\/colpali-se-passer-docr-pour-la-recherche\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/askem.eu\\\/2026\\\/05\\\/11\\\/colpali-se-passer-docr-pour-la-recherche\\\/#primaryimage\",\"url\":\"https:\\/\\/askem.eu\\/wp-content\\/uploads\\/2026\\/05\\/sujet-askem-2026-05-08.png\",\"contentUrl\":\"https:\\/\\/askem.eu\\/wp-content\\/uploads\\/2026\\/05\\/sujet-askem-2026-05-08.png\",\"width\":1600,\"height\":1200},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/askem.eu\\\/2026\\\/05\\\/11\\\/colpali-se-passer-docr-pour-la-recherche\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Accueil\",\"item\":\"https:\\\/\\\/askem.eu\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"ColPali, se passer d&rsquo;OCR pour la recherche\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/askem.eu\\\/#website\",\"url\":\"https:\\\/\\\/askem.eu\\\/\",\"name\":\"askem\",\"description\":\"\",\"publisher\":{\"@id\":\"https:\\\/\\\/askem.eu\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/askem.eu\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/askem.eu\\\/#organization\",\"name\":\"Askem\",\"url\":\"https:\\\/\\\/askem.eu\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/askem.eu\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\/\\/mlpi0fxo3sth.i.optimole.com\\/cb:3obA.c61\\/w:760\\/h:480\\/q:mauto\\/f:best\\/https:\\/\\/askem.eu\\/wp-content\\/uploads\\/2020\\/10\\/logoGalaxieAskem3.png\",\"contentUrl\":\"https:\\/\\/mlpi0fxo3sth.i.optimole.com\\/cb:3obA.c61\\/w:760\\/h:480\\/q:mauto\\/f:best\\/https:\\/\\/askem.eu\\/wp-content\\/uploads\\/2020\\/10\\/logoGalaxieAskem3.png\",\"width\":760,\"height\":480,\"caption\":\"Askem\"},\"image\":{\"@id\":\"https:\\\/\\\/askem.eu\\\/#\\\/schema\\\/logo\\\/image\\\/\"},\"sameAs\":[\"https:\\\/\\\/fb.me\\\/askem.eu\",\"https:\\\/\\\/linkedin.com\\\/company\\\/askem-eu\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/askem.eu\\\/#\\\/schema\\\/person\\\/8bbee74ab9a977d56bf4826662e9d2e9\",\"name\":\"askemadmin\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/a202f744ee3a4b6fdbe2ceb57fd84c72559337791a276662270d8d2fb7842e3f?s=96&d=mm&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/a202f744ee3a4b6fdbe2ceb57fd84c72559337791a276662270d8d2fb7842e3f?s=96&d=mm&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/a202f744ee3a4b6fdbe2ceb57fd84c72559337791a276662270d8d2fb7842e3f?s=96&d=mm&r=g\",\"caption\":\"askemadmin\"},\"sameAs\":[\"https:\\\/\\\/askem.eu\"]}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"ColPali, se passer d'OCR pour la recherche - askem","description":"ASKEM BUREAU D'\u00c9TUDES ET DE FORMATION NUM\u00c9RIQUE. Nous vous assistons dans la transformation num\u00e9rique de vos outils, services et organisations tout en pla\u00e7ant l\u2019humain au c\u0153ur de notre d\u00e9marche d\u2019accompagnement.","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:\/\/askem.eu\/en\/2026\/05\/11\/colpali-se-passer-docr-pour-la-recherche\/","og_locale":"en_US","og_type":"article","og_title":"ColPali, se passer d'OCR pour la recherche - askem","og_description":"ASKEM BUREAU D'\u00c9TUDES ET DE FORMATION NUM\u00c9RIQUE. Nous vous assistons dans la transformation num\u00e9rique de vos outils, services et organisations tout en pla\u00e7ant l\u2019humain au c\u0153ur de notre d\u00e9marche d\u2019accompagnement.","og_url":"https:\/\/askem.eu\/en\/2026\/05\/11\/colpali-se-passer-docr-pour-la-recherche\/","og_site_name":"askem","article_publisher":"https:\/\/fb.me\/askem.eu","article_published_time":"2026-05-11T09:21:38+00:00","article_modified_time":"2026-05-11T09:21:44+00:00","og_image":[{"width":1600,"height":1200,"url":"https:\/\/mlpi0fxo3sth.i.optimole.com\/cb:3obA.c61\/w:auto\/h:auto\/q:mauto\/f:best\/https:\/\/askem.eu\/wp-content\/uploads\/2026\/05\/sujet-askem-2026-05-08.png","type":"image\/png"}],"author":"askemadmin","twitter_card":"summary_large_image","twitter_misc":{"Written by":"askemadmin","Est. reading time":"6 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/askem.eu\/2026\/05\/11\/colpali-se-passer-docr-pour-la-recherche\/#article","isPartOf":{"@id":"https:\/\/askem.eu\/2026\/05\/11\/colpali-se-passer-docr-pour-la-recherche\/"},"author":{"name":"askemadmin","@id":"https:\/\/askem.eu\/#\/schema\/person\/8bbee74ab9a977d56bf4826662e9d2e9"},"headline":"ColPali, se passer d&rsquo;OCR pour la recherche","datePublished":"2026-05-11T09:21:38+00:00","dateModified":"2026-05-11T09:21:44+00:00","mainEntityOfPage":{"@id":"https:\/\/askem.eu\/2026\/05\/11\/colpali-se-passer-docr-pour-la-recherche\/"},"wordCount":1238,"commentCount":0,"publisher":{"@id":"https:\/\/askem.eu\/#organization"},"image":{"@id":"https:\/\/askem.eu\/2026\/05\/11\/colpali-se-passer-docr-pour-la-recherche\/#primaryimage"},"thumbnailUrl":"https:\/\/mlpi0fxo3sth.i.optimole.com\/cb:3obA.c61\/w:auto\/h:auto\/q:mauto\/f:best\/https:\/\/askem.eu\/wp-content\/uploads\/2026\/05\/sujet-askem-2026-05-08.png","articleSection":["AI"],"inLanguage":"en-US","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/askem.eu\/2026\/05\/11\/colpali-se-passer-docr-pour-la-recherche\/#respond"]}]},{"@type":"WebPage","@id":"https:\/\/askem.eu\/2026\/05\/11\/colpali-se-passer-docr-pour-la-recherche\/","url":"https:\/\/askem.eu\/2026\/05\/11\/colpali-se-passer-docr-pour-la-recherche\/","name":"ColPali, se passer d'OCR pour la recherche - askem","isPartOf":{"@id":"https:\/\/askem.eu\/#website"},"primaryImageOfPage":{"@id":"https:\/\/askem.eu\/2026\/05\/11\/colpali-se-passer-docr-pour-la-recherche\/#primaryimage"},"image":{"@id":"https:\/\/askem.eu\/2026\/05\/11\/colpali-se-passer-docr-pour-la-recherche\/#primaryimage"},"thumbnailUrl":"https:\/\/mlpi0fxo3sth.i.optimole.com\/cb:3obA.c61\/w:auto\/h:auto\/q:mauto\/f:best\/https:\/\/askem.eu\/wp-content\/uploads\/2026\/05\/sujet-askem-2026-05-08.png","datePublished":"2026-05-11T09:21:38+00:00","dateModified":"2026-05-11T09:21:44+00:00","description":"ASKEM BUREAU D'\u00c9TUDES ET DE FORMATION NUM\u00c9RIQUE. Nous vous assistons dans la transformation num\u00e9rique de vos outils, services et organisations tout en pla\u00e7ant l\u2019humain au c\u0153ur de notre d\u00e9marche d\u2019accompagnement.","breadcrumb":{"@id":"https:\/\/askem.eu\/2026\/05\/11\/colpali-se-passer-docr-pour-la-recherche\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/askem.eu\/2026\/05\/11\/colpali-se-passer-docr-pour-la-recherche\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/askem.eu\/2026\/05\/11\/colpali-se-passer-docr-pour-la-recherche\/#primaryimage","url":"https:\/\/mlpi0fxo3sth.i.optimole.com\/cb:3obA.c61\/w:auto\/h:auto\/q:mauto\/f:best\/https:\/\/askem.eu\/wp-content\/uploads\/2026\/05\/sujet-askem-2026-05-08.png","contentUrl":"https:\/\/mlpi0fxo3sth.i.optimole.com\/cb:3obA.c61\/w:auto\/h:auto\/q:mauto\/f:best\/https:\/\/askem.eu\/wp-content\/uploads\/2026\/05\/sujet-askem-2026-05-08.png","width":1600,"height":1200},{"@type":"BreadcrumbList","@id":"https:\/\/askem.eu\/2026\/05\/11\/colpali-se-passer-docr-pour-la-recherche\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Accueil","item":"https:\/\/askem.eu\/"},{"@type":"ListItem","position":2,"name":"ColPali, se passer d&rsquo;OCR pour la recherche"}]},{"@type":"WebSite","@id":"https:\/\/askem.eu\/#website","url":"https:\/\/askem.eu\/","name":"askem","description":"","publisher":{"@id":"https:\/\/askem.eu\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/askem.eu\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/askem.eu\/#organization","name":"Askem","url":"https:\/\/askem.eu\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/askem.eu\/#\/schema\/logo\/image\/","url":"https:\/\/mlpi0fxo3sth.i.optimole.com\/cb:3obA.c61\/w:760\/h:480\/q:mauto\/f:best\/https:\/\/askem.eu\/wp-content\/uploads\/2020\/10\/logoGalaxieAskem3.png","contentUrl":"https:\/\/mlpi0fxo3sth.i.optimole.com\/cb:3obA.c61\/w:760\/h:480\/q:mauto\/f:best\/https:\/\/askem.eu\/wp-content\/uploads\/2020\/10\/logoGalaxieAskem3.png","width":760,"height":480,"caption":"Askem"},"image":{"@id":"https:\/\/askem.eu\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/fb.me\/askem.eu","https:\/\/linkedin.com\/company\/askem-eu"]},{"@type":"Person","@id":"https:\/\/askem.eu\/#\/schema\/person\/8bbee74ab9a977d56bf4826662e9d2e9","name":"askemadmin","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/secure.gravatar.com\/avatar\/a202f744ee3a4b6fdbe2ceb57fd84c72559337791a276662270d8d2fb7842e3f?s=96&d=mm&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/a202f744ee3a4b6fdbe2ceb57fd84c72559337791a276662270d8d2fb7842e3f?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/a202f744ee3a4b6fdbe2ceb57fd84c72559337791a276662270d8d2fb7842e3f?s=96&d=mm&r=g","caption":"askemadmin"},"sameAs":["https:\/\/askem.eu"]}]}},"_links":{"self":[{"href":"https:\/\/askem.eu\/en\/wp-json\/wp\/v2\/posts\/2314","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/askem.eu\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/askem.eu\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/askem.eu\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/askem.eu\/en\/wp-json\/wp\/v2\/comments?post=2314"}],"version-history":[{"count":1,"href":"https:\/\/askem.eu\/en\/wp-json\/wp\/v2\/posts\/2314\/revisions"}],"predecessor-version":[{"id":2316,"href":"https:\/\/askem.eu\/en\/wp-json\/wp\/v2\/posts\/2314\/revisions\/2316"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/askem.eu\/en\/wp-json\/wp\/v2\/media\/2315"}],"wp:attachment":[{"href":"https:\/\/askem.eu\/en\/wp-json\/wp\/v2\/media?parent=2314"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/askem.eu\/en\/wp-json\/wp\/v2\/categories?post=2314"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/askem.eu\/en\/wp-json\/wp\/v2\/tags?post=2314"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}