Abstract: We introduce WildVideo, an open-world benchmark dataset designed to address how to assess hallucination of Large Multi-modal Models (LMMs) for understanding video-language interaction in the ...
Abstract: While abundant research has been conducted on improving high-level visual understanding and reasoning capabilities of large multimodal models (LMMs), their image quality assessment (IQA) ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results