ELARKHNU browse.title.page

Зараз показуємо 1 - 1 з 1

Practice Implementation of Neural Network Model BART-Large-CNN for Text Annotation
(2024) Mazurets, O.; Molchanova, M.; Klimenko, V.; Prosvitliuk, M; Мазурець, Олександр Вікторович
The scheme of the method of annotating works of art was described, which works by converting input data in the form of text for annotating a work of art, a trained machine learning model, and desired annotation parameters into output data in the form of an annotation and a numerical evaluation of the quality of the annotation and is intended for automated annotation creation. We also present the neural network architecture of the machine learning model, which is the input to the proposed method of annotating works of art. This neural network model belongs to the «transformers» type and is currently one of the most powerful text generation models. A practical implementation of the method of annotating works of art has been created and the main purposes of the software components of the intelligent system for annotating works of art have been described.