vix.ing · top · new · best · stats · spec

LLaVA-OneVision: Easy Visual Task Transfer

2024/08/06 by Bo Li, Yuanhan Zhang, Li, Bo +18 · 588 citations
Computer Science · Engineering · #Gaze Tracking and Assistive Technology #Teleoperation and Haptic Systems

paper · pdf · doi:10.48550/arxiv.2408.03326

Abstract

We present LLaVA-OneVision, a family of open large multimodal models (LMMs) developed by consolidating our insights into data, models, and visual representations in the LLaVA-NeXT blog series. Our experimental results demonstrate that LLaVA-OneVision is the first single model that can simultaneously push the performance boundaries of open LMMs in three important computer vision scenarios: single-image, multi-image, and video scenarios. Importantly, the design of LLaVA-OneVision allows strong transfer learning across different modalities/scenarios, yielding new emerging capabilities. In particular, strong video understanding and cross-scenario capabilities are demonstrated through task transfer from images to videos.

Cited by

Related