Lightweight visual accessibility LLaVA architecture

Abstract At present, many models focus on the field of blind barrier-free recognition, but most of them are simple visual recognition. Multimodal visual language models face the challenges of high computational cost and poor real-time performance in blind barrier-free recognition tasks, which limits...

Popoln opis

Bibliografske podrobnosti

Principais autores:	Zhiyin Han, Xiaoqun Liu, Juan Hao
Format:	Artigo
Jezik:	Inglês
Izdano:	Nature Portfolio 2025-11-01
Serija:	Scientific Reports
Teme:	Multimodal architecture Visual accessibility LLaVA Sparse experts Perceptual weights
Online dostop:	https://doi.org/10.1038/s41598-025-23023-w
Oznake:	Označite Brez oznak, prvi označite!