Connexion
Idée - Document Processing
Japanese Document Question Answering (JDocQA) Dataset for Generative Language Models
JDocQA is a large-scale Japanese document question answering dataset that requires understanding of both textual and visual information to answer questions.
ViTLP: Visually Guided Generative Text-Layout Pre-training for Document Intelligence
Visually guided generative text-layout pre-training enhances document understanding by optimizing hierarchical language and layout modeling objectives.
Visually Guided Generative Text-Layout Pre-training for Document Intelligence
ViTLP proposes visually guided generative text-layout pre-training to enhance document understanding by optimizing hierarchical language and layout modeling objectives.
ANLS* - A Universal Document Processing Metric for Generative Large Language Models
Introducing ANLS*, a versatile metric for evaluating generative models in document processing tasks.
Produits | Ressources
© 2024 by Linnk AI