{"article":{"annotation":{"article_id":"alphaedge-releases-french-sparse-embedding-model-october-2026","available_at":"2026-10-15T19:00:52.803533Z","geography_revision":"ef6ee0010dda5fe7b240fd514749b43e214c57da73222d2e40ec83b276dd0977","id":"7d3689cabe55567830aad36c8aebffa505cb55c500896d3562af813ff21c1532","input_sha256":"b0f4c63f9545b6354cb1a32ef1f4da61431b6c9b68e67d7f3b0bb677d7840a0c","limitations":["Rule-derived subjects are candidates, not verified editorial labels or calibrated model probabilities.","Coverage is distinct from event location. Unknown country/city/coordinates remain absent; no automatic geocoding.","Article geometries keep their stated precision and methodology; reproducing them is not independent coordinate verification."],"method":"controlled-rules-v1","places":[],"rules_sha256":"fb5ad3c27c91cb840f7a0a05f7f64a5cb26ad0374409ff9457f277c7092e22b4","schema_version":1,"subjects":[{"concept_id":"medtop:20001298","confidence":null,"confidence_method":null,"evidence":[{"field":"en.title+en.summary","language":"en","note":"Subject suggested from the title or summary: AI","source_urls":["https://huggingface.co/blog/alphaedge-ai/french-sparseup"]}],"method":"controlled-rules-v1","rule_id":"artificial-intelligence","status":"candidate"}],"taxonomy_revision":"1f2db31d811fca07e99b16d402c10e2cc2965c9df1c7b41d9be348b40a24e7fb","taxonomy_version":"IPTC Media Topics 2026-Q2"},"api_url":"https://newsverum.com/api/v1/articles/alphaedge-releases-french-sparse-embedding-model-october-2026?lang=fr","category":"technology","content":{"key_points":[],"paragraphs":["AlphaEdge AI a publié sur Hugging Face Sparseup-fr, une reproduction en français de l’approche d’embeddings textuels sparse de Linkup. L’équipe indique avoir utilisé la base multilingue mLateOn-unsupervised et une tête de modèle de langue masqué mmBERT-base réduite à un vocabulaire français ; l’entraînement a duré environ 55 heures sur un seul GPU RTX 6000 Pro.","Sur le benchmark NanoBEIR-fr nettoyé de 13 jeux de données, Sparseup-fr arrive en tête sur huit, le modèle sparse multilingue comparé d’OpenSearch sur trois et SPLADE CamemBERT sur deux. L’équipe dit avoir recherché d’éventuelles fuites entre données d’entraînement et benchmark et supprimé des lignes suspectes, sans examiner les possibles problèmes d’annotation. La publication comprend le modèle et le benchmark nettoyé ; les résultats sont l’évaluation des auteurs du modèle."],"sections":[]},"corrections":[],"country":null,"data":{"blocks":[{"description":{"ar":"عدد مجموعات البيانات التي تصدرها كل نموذج في معيار NanoBEIR الفرنسي المنقح، المكوّن من 13 مجموعة.","de":"Zahl der Datensätze, bei denen jedes Modell im bereinigten französischen NanoBEIR-Benchmark mit 13 Datensätzen vorn lag.","en":"Number of datasets led by each model on the cleaned 13-dataset French NanoBEIR benchmark.","es":"Número de conjuntos de datos en los que cada modelo quedó primero en la prueba francesa NanoBEIR depurada, con 13 conjuntos.","fr":"Nombre de jeux de données où chaque modèle arrive en tête sur le benchmark français NanoBEIR nettoyé, qui compte 13 jeux.","pt":"Número de conjuntos de dados liderados por cada modelo no benchmark francês NanoBEIR depurado, com 13 conjuntos.","tr":"Temizlenmiş 13 veri kümeli Fransızca NanoBEIR kıyaslamasında her modelin en iyi olduğu veri kümesi sayısı."},"id":"nanobeir-fr-sparse-model-leads","items":[{"basis":"observed","label":{"ar":"مجموعات البيانات التي تصدرها Sparseup-fr","de":"Datensätze mit Sparseup-fr an der Spitze","en":"Datasets led by Sparseup-fr","es":"Conjuntos liderados por Sparseup-fr","fr":"Jeux de données en tête pour Sparseup-fr","pt":"Conjuntos liderados pelo Sparseup-fr","tr":"Sparseup-fr'nin ilk sırada olduğu veri kümesi"},"measure":{"aggregation":"stock","frequency":"irregular","geography":"world","id":"nanobeir-fr-dataset-leads-sparseup-fr","periodRole":"announcement-date","unitCode":"count"},"period":"2026-10-12","unit":"count","value":8},{"basis":"observed","label":{"ar":"مجموعات البيانات التي يتصدرها نموذج OpenSearch متعدد اللغات","de":"Datensätze mit dem mehrsprachigen OpenSearch-Sparse-Modell an der Spitze","en":"Datasets led by the OpenSearch multilingual sparse model","es":"Conjuntos liderados por el modelo disperso multilingüe de OpenSearch","fr":"Jeux de données en tête pour le modèle sparse multilingue d’OpenSearch","pt":"Conjuntos liderados pelo modelo esparso multilingue da OpenSearch","tr":"OpenSearch multilingual sparse modelinin ilk sırada olduğu veri kümesi"},"measure":{"aggregation":"stock","frequency":"irregular","geography":"world","id":"nanobeir-fr-dataset-leads-opensearch","periodRole":"announcement-date","unitCode":"count"},"period":"2026-10-12","unit":"count","value":3},{"basis":"observed","label":{"ar":"مجموعات البيانات التي يتصدرها نموذج SPLADE CamemBERT","de":"Datensätze mit dem SPLADE-CamemBERT-Modell an der Spitze","en":"Datasets led by the SPLADE CamemBERT model","es":"Conjuntos liderados por el modelo SPLADE CamemBERT","fr":"Jeux de données en tête pour le modèle SPLADE CamemBERT","pt":"Conjuntos liderados pelo modelo SPLADE CamemBERT","tr":"SPLADE CamemBERT modelinin ilk sırada olduğu veri kümesi"},"measure":{"aggregation":"stock","frequency":"irregular","geography":"world","id":"nanobeir-fr-dataset-leads-splade","periodRole":"announcement-date","unitCode":"count"},"period":"2026-10-12","unit":"count","value":2}],"methodology":{"ar":"حُسبت الأعداد من جدول نتائج NanoBEIR-fr المنقح في منشور AlphaEdge AI بتاريخ 12 أكتوبر 2026. قال المؤلفون إنهم أزالوا تسرب البيانات والصفوف المشبوهة لكنهم لم يفحصوا مشكلات الوسم المحتملة. الأعداد ليست تقييماً مستقلاً.","de":"Die Zahlen wurden aus der bereinigten NanoBEIR-fr-Ergebnistabelle im Beitrag von AlphaEdge AI vom 12. Oktober 2026 gezählt. Die Autoren erklärten, Datenlecks und verdächtige Zeilen entfernt, mögliche Annotationsprobleme aber nicht untersucht zu haben. Es handelt sich nicht um eine unabhängige Auswertung.","en":"Counted from the cleaned NanoBEIR-fr results table in AlphaEdge AI's 12 October 2026 post. The authors said they removed data leakage and suspicious rows but did not investigate possible annotation problems. Counts are not an independent evaluation.","es":"Las cifras se contaron en la tabla de resultados NanoBEIR-fr depurada publicada por AlphaEdge AI el 12 de octubre de 2026. Los autores dijeron haber eliminado filtraciones de datos y filas sospechosas, pero no investigaron posibles problemas de anotación. No es una evaluación independiente.","fr":"Les nombres sont comptés dans le tableau NanoBEIR-fr nettoyé publié par AlphaEdge AI le 12 octobre 2026. Les auteurs disent avoir retiré les fuites de données et les lignes suspectes, sans examiner d’éventuels problèmes d’annotation. Il ne s’agit pas d’une évaluation indépendante.","pt":"Os valores foram contados na tabela de resultados NanoBEIR-fr depurada publicada pela AlphaEdge AI em 12 de outubro de 2026. Os autores disseram ter removido fugas de dados e linhas suspeitas, mas não investigaram possíveis problemas de anotação. Não é uma avaliação independente.","tr":"AlphaEdge AI'nın 12 Ekim 2026 tarihli yazısında verdiği temizlenmiş NanoBEIR-fr sonuç tablosundan sayılmıştır. Yazarlar veri sızıntısı ve şüpheli satırları temizlediklerini, olası etiketleme sorunlarını ise incelemediklerini belirtiyor. Sayılar bağımsız değerlendirme değildir."},"sourceUrls":["https://huggingface.co/blog/alphaedge-ai/french-sparseup"],"title":{"ar":"النماذج المتناثرة المتقدمة على NanoBEIR-fr","de":"Führende Sparse-Modelle auf NanoBEIR-fr","en":"Leading sparse models on NanoBEIR-fr","es":"Modelos dispersos líderes en NanoBEIR-fr","fr":"Modèles épars en tête sur NanoBEIR-fr","pt":"Modelos esparsos líderes no NanoBEIR-fr","tr":"NanoBEIR-fr'de önde gelen seyrek modeller"},"type":"metrics","updatedAt":"2026-10-15T18:50:44+00:00"}],"version":1},"data_downloads":[{"formats":{"csv":"https://newsverum.com/api/v1/articles/alphaedge-releases-french-sparse-embedding-model-october-2026/data/nanobeir-fr-sparse-model-leads?lang=fr&format=csv","json":"https://newsverum.com/api/v1/articles/alphaedge-releases-french-sparse-embedding-model-october-2026/data/nanobeir-fr-sparse-model-leads?lang=fr&format=json"},"id":"nanobeir-fr-sparse-model-leads"}],"data_types":["metrics"],"editorial_illustration":{"alt":"Illustration conceptuelle de la technologie, des processeurs et des informations connectées.","asset_id":"newsverum-v1-technology","avif_variants":[{"height":320,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-480.avif","width":480},{"height":480,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-720.avif","width":720},{"height":640,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-960.avif","width":960},{"height":800,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1200.avif","width":1200},{"height":1024,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1536.avif","width":1536}],"credit":"Newsverum","documentary":false,"height":1024,"kind":"symbolic-editorial-illustration","language":"fr","scope":"topic","url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1536.webp","variants":[{"height":320,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-480.webp","width":480},{"height":640,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-960.webp","width":960},{"height":1024,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1536.webp","width":1536}],"width":1536},"event_id":"event-70cbde2b-ba3d-5e3e-b9aa-7a69e76f07b9","event_period":"2026-10-12","event_period_basis":"event-or-reference-period","id":"alphaedge-releases-french-sparse-embedding-model-october-2026","image":null,"importance":0.6,"language":"fr","modified_at":"2026-10-15T19:00:44+00:00","publication_basis":"publication-record","published_at":"2026-10-15T19:00:44+00:00","reading_minutes":1,"region":"world","relations":[],"section_illustration":{"alt":"Illustration conceptuelle de la technologie, des processeurs et des informations connectées.","asset_id":"newsverum-v1-technology","avif_variants":[{"height":320,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-480.avif","width":480},{"height":480,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-720.avif","width":720},{"height":640,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-960.avif","width":960},{"height":800,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1200.avif","width":1200},{"height":1024,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1536.avif","width":1536}],"credit":"Newsverum","documentary":false,"height":1024,"kind":"symbolic-editorial-illustration","language":"fr","scope":"section","url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1536.webp","variants":[{"height":320,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-480.webp","width":480},{"height":640,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-960.webp","width":960},{"height":1024,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1536.webp","width":1536}],"width":1536},"slug":"alphaedge-releases-french-sparse-embedding-model-october-2026","sources":[{"checkedAt":"2026-10-15","name":"AlphaEdge AI","publishedAt":"2026-10-12","url":"https://huggingface.co/blog/alphaedge-ai/french-sparseup"}],"summary":"AlphaEdge AI a publié Sparseup-fr, une adaptation française de la méthode open source de Linkup, avec un benchmark NanoBEIR-fr nettoyé.","text":"AlphaEdge AI a publié Sparseup-fr, une adaptation française de la méthode open source de Linkup, avec un benchmark NanoBEIR-fr nettoyé.\n\nAlphaEdge AI a publié sur Hugging Face Sparseup-fr, une reproduction en français de l’approche d’embeddings textuels sparse de Linkup. L’équipe indique avoir utilisé la base multilingue mLateOn-unsupervised et une tête de modèle de langue masqué mmBERT-base réduite à un vocabulaire français ; l’entraînement a duré environ 55 heures sur un seul GPU RTX 6000 Pro.\n\nSur le benchmark NanoBEIR-fr nettoyé de 13 jeux de données, Sparseup-fr arrive en tête sur huit, le modèle sparse multilingue comparé d’OpenSearch sur trois et SPLADE CamemBERT sur deux. L’équipe dit avoir recherché d’éventuelles fuites entre données d’entraînement et benchmark et supprimé des lignes suspectes, sans examiner les possibles problèmes d’annotation. La publication comprend le modèle et le benchmark nettoyé ; les résultats sont l’évaluation des auteurs du modèle.","title":"AlphaEdge AI publie un modèle d’embeddings textuels sparse en français","topics":[],"translations":[{"api_url":"https://newsverum.com/api/v1/articles/alphaedge-releases-french-sparse-embedding-model-october-2026?lang=tr","language":"tr","url":"https://newsverum.com/tr/news/alphaedge-releases-french-sparse-embedding-model-october-2026"},{"api_url":"https://newsverum.com/api/v1/articles/alphaedge-releases-french-sparse-embedding-model-october-2026?lang=en","language":"en","url":"https://newsverum.com/en/news/alphaedge-releases-french-sparse-embedding-model-october-2026"},{"api_url":"https://newsverum.com/api/v1/articles/alphaedge-releases-french-sparse-embedding-model-october-2026?lang=ar","language":"ar","url":"https://newsverum.com/ar/news/alphaedge-releases-french-sparse-embedding-model-october-2026"},{"api_url":"https://newsverum.com/api/v1/articles/alphaedge-releases-french-sparse-embedding-model-october-2026?lang=fr","language":"fr","url":"https://newsverum.com/fr/news/alphaedge-releases-french-sparse-embedding-model-october-2026"},{"api_url":"https://newsverum.com/api/v1/articles/alphaedge-releases-french-sparse-embedding-model-october-2026?lang=es","language":"es","url":"https://newsverum.com/es/news/alphaedge-releases-french-sparse-embedding-model-october-2026"},{"api_url":"https://newsverum.com/api/v1/articles/alphaedge-releases-french-sparse-embedding-model-october-2026?lang=de","language":"de","url":"https://newsverum.com/de/news/alphaedge-releases-french-sparse-embedding-model-october-2026"},{"api_url":"https://newsverum.com/api/v1/articles/alphaedge-releases-french-sparse-embedding-model-october-2026?lang=pt","language":"pt","url":"https://newsverum.com/pt/news/alphaedge-releases-french-sparse-embedding-model-october-2026"}],"url":"https://newsverum.com/fr/news/alphaedge-releases-french-sparse-embedding-model-october-2026"},"version":"1"}