List your company

Multimodal Visual Question Answering on SLAKE

32.62AccuracySPINE
8.107214.471120.83527.1989Nov 22, 2025

Evaluation results

MethodLinks
SPINEBase Model=Qwen2.5-VL-3B-Instruct2025.1132.62
TTRLBase Model=Qwen2.5-VL-3B-Instruct2025.1130
No adaptationBase Model=Qwen2.5-VL-3B-Instruct2025.1126.17
Self-ConsistencyBase Model=Qwen2.5-VL-3B-Instruct2025.1125.84
SEALONGBase Model=Qwen2.5-VL-3B-Instruct2025.1112.32
LMSIBase Model=Qwen2.5-VL-3B-Instruct2025.119.05