{"id":846,"date":"2023-09-11T17:36:47","date_gmt":"2023-09-11T17:36:47","guid":{"rendered":"https:\/\/sciencesetrecherches.eu\/?p=846"},"modified":"2023-09-11T17:36:48","modified_gmt":"2023-09-11T17:36:48","slug":"comment-interagir-avec-les-modeles-existant","status":"publish","type":"post","link":"https:\/\/sciencesetrecherches.eu\/?p=846","title":{"rendered":"Comment interagir avec les mod\u00e8les existant"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\">voici la proc\u00e9dure pour interagir avec des mod\u00e8les que vous auriez pu t\u00e9l\u00e9charger :<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><\/p>\n\n\n\n<div id=\"wp-block-themeisle-blocks-circle-counter-0bad26a6\" data-percentage=\"50\" data-duration=\"2\" data-height=\"100\" data-stroke-width=\"10\" class=\"wp-block-themeisle-blocks-circle-counter\"><div class=\"wp-block-themeisle-blocks-circle-counter-title__area\"><span class=\"wp-block-themeisle-blocks-circle-counter-title__value\">Skill<\/span><\/div><div class=\"wp-block-themeisle-blocks-circle-counter__bar\"><\/div><\/div>\n\n\n\n<p class=\"wp-block-paragraph\">Voici un guide rapide pour d\u00e9marrer avec Hugging Face&#8217;s Transformers:<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">1. Installation:<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">Installez la biblioth\u00e8que via pip:<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>pip install transformers<\/code><\/pre>\n\n\n\n<h3 class=\"wp-block-heading\">2. Utilisation de base:<\/h3>\n\n\n\n<h4 class=\"wp-block-heading\">Charger un mod\u00e8le et un tokenizer:<\/h4>\n\n\n\n<pre class=\"wp-block-code\"><code>from transformers import AutoModel, AutoTokenizer\n\n# Pour un mod\u00e8le fran\u00e7ais, par exemple CamemBERT\nmodel_name = \"camembert-base\"\nmodel = AutoModel.from_pretrained(model_name)\ntokenizer = AutoTokenizer.from_pretrained(model_name)<\/code><\/pre>\n\n\n\n<h4 class=\"wp-block-heading\">Tokenisation et d\u00e9tokenisation:<\/h4>\n\n\n\n<pre class=\"wp-block-code\"><code>text = \"Bonjour le monde!\"\ntokens = tokenizer.tokenize(text)\nprint(tokens)\n\n# Convertir les tokens en IDs et vice versa\ninput_ids = tokenizer.convert_tokens_to_ids(tokens)\nprint(input_ids)\n\noriginal_tokens = tokenizer.convert_ids_to_tokens(input_ids)\nprint(original_tokens)<\/code><\/pre>\n\n\n\n<h4 class=\"wp-block-heading\">Utilisation du mod\u00e8le:<\/h4>\n\n\n\n<pre class=\"wp-block-code\"><code># Convertir le texte en entr\u00e9es de mod\u00e8le\ninputs = tokenizer(text, return_tensors=\"pt\")\n\n# Obtenir les embeddings du texte\nwith torch.no_grad():\n    embeddings = model(**inputs).last_hidden_state\nprint(embeddings)<\/code><\/pre>\n\n\n\n<h3 class=\"wp-block-heading\">3. T\u00e2ches sp\u00e9cifiques:<\/h3>\n\n\n\n<h4 class=\"wp-block-heading\">Classification de texte:<\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">Si vous avez un mod\u00e8le sp\u00e9cifique pour la classification, vous pouvez le charger et l&#8217;utiliser pour pr\u00e9dire des classes pour des textes donn\u00e9s.<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>from transformers import CamembertForSequenceClassification\n\nmodel = CamembertForSequenceClassification.from_pretrained(\"path_to_model\")\npredictions = model(**inputs).logits<\/code><\/pre>\n\n\n\n<h4 class=\"wp-block-heading\">G\u00e9n\u00e9ration de texte:<\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">Pour les mod\u00e8les de g\u00e9n\u00e9ration comme GPT-2, vous pouvez les utiliser pour g\u00e9n\u00e9rer du texte en fran\u00e7ais.<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>from transformers import GPT2LMHeadModel, GPT2Tokenizer\n\nmodel_name = \"gpt2-medium\"\nmodel = GPT2LMHeadModel.from_pretrained(model_name)\ntokenizer = GPT2Tokenizer.from_pretrained(model_name)\n\ngenerated = model.generate(**inputs)\ngenerated_text = tokenizer.decode(generated&#91;0], skip_special_tokens=True)\nprint(generated_text)<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">Ceci n&#8217;est qu&#8217;une introduction rapide \u00e0 Hugging Face&#8217;s Transformers. La biblioth\u00e8que est vaste et offre de nombreuses fonctionnalit\u00e9s avanc\u00e9es. Je vous recommande de consulter la <a href=\"https:\/\/huggingface.co\/transformers\/\">documentation officielle<\/a> pour plus de d\u00e9tails et d&#8217;exemples.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>voici la proc\u00e9dure pour interagir avec des mod\u00e8les que vous auriez pu t\u00e9l\u00e9charger : Voici un guide rapide pour d\u00e9marrer avec Hugging Face&#8217;s Transformers: 1. Installation: Installez la biblioth\u00e8que via pip: 2. Utilisation de base: Charger un mod\u00e8le et un tokenizer: Tokenisation et d\u00e9tokenisation: Utilisation du mod\u00e8le: 3. T\u00e2ches sp\u00e9cifiques: Classification de texte: Si vous [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":true,"template":"","format":"standard","meta":{"_themeisle_gutenberg_block_has_review":false,"footnotes":""},"categories":[31,75,29,76],"tags":[34,78,32,79,77],"series":[],"class_list":["post-846","post","type-post","status-publish","format-standard","hentry","category-gpt","category-hugging","category-ia","category-transformers","tag-gpt","tag-hugging-face","tag-ia","tag-modeles-ia","tag-transformers"],"_links":{"self":[{"href":"https:\/\/sciencesetrecherches.eu\/index.php?rest_route=\/wp\/v2\/posts\/846","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/sciencesetrecherches.eu\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/sciencesetrecherches.eu\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/sciencesetrecherches.eu\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/sciencesetrecherches.eu\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=846"}],"version-history":[{"count":1,"href":"https:\/\/sciencesetrecherches.eu\/index.php?rest_route=\/wp\/v2\/posts\/846\/revisions"}],"predecessor-version":[{"id":847,"href":"https:\/\/sciencesetrecherches.eu\/index.php?rest_route=\/wp\/v2\/posts\/846\/revisions\/847"}],"wp:attachment":[{"href":"https:\/\/sciencesetrecherches.eu\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=846"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/sciencesetrecherches.eu\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=846"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/sciencesetrecherches.eu\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=846"},{"taxonomy":"series","embeddable":true,"href":"https:\/\/sciencesetrecherches.eu\/index.php?rest_route=%2Fwp%2Fv2%2Fseries&post=846"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}