SOTAVerified

LLaVA-OneVision: Easy Visual Task Transfer

2024-08-06Code Available0· sign in to hype

Bo Li, Yuanhan Zhang, Dong Guo, Renrui Zhang, Feng Li, Hao Zhang, Kaichen Zhang, Peiyuan Zhang, Yanwei Li, Ziwei Liu, Chunyuan Li

Code Available — Be the first to reproduce this paper.

Reproduce

Code

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.

Tasks

Benchmark Results

DatasetModelMetricClaimedVerifiedStatus
VinogroundLLaVA-OneVision-Qwen2-7BText Score41.6Unverified
VinogroundLLaVA-OneVision-Qwen2-72BText Score48.4Unverified

Reproductions