List your company

Visual Question Answering (Perception) on CV-Bench, BLINK, and RealWorldQA

60.2CV-Bench ScoreLATTE
50.21652.80855.457.992Dec 7, 2024

Evaluation results

MethodLinks
LATTELanguage / Vision=Qwen2-7B/SigLIP, Starting checkpoint=LLaVA-OV Stage 1.52024.1260.252.661.158
LATTELanguage / Vision=LLaMA3-8B/CLIP, Starting checkpoint=Mantis Pretrained (Configuration 2)2024.1257.247.853.752.9
LATTELanguage / Vision=LLaMA3-8B/CLIP, Starting checkpoint=Mantis Pretrained2024.1256.949.651.152.6
Vanilla ITLanguage / Vision=Qwen2-7B/SigLIP, Starting checkpoint=LLaVA-OV Stage 1.52024.1256.850.357.855
Vanilla ITLanguage / Vision=LLaMA3-8B/CLIP, Starting checkpoint=Mantis Pretrained2024.1252.645.852.350.2
Vanilla ITLanguage / Vision=LLaMA3-8B/CLIP, Starting checkpoint=Mantis Pretrained (Configuration 2)2024.1252.343.751.849.3
LATTELanguage / Vision=LLaMA3-8B/SigLIP, Starting checkpoint=Mantis Instruct-tuned2024.1251.747.356.151.7
Vanilla ITLanguage / Vision=LLaMA3-8B/SigLIP, Starting checkpoint=Mantis Instruct-tuned2024.1250.646.754.850.7