List your company

Visual Question Answering on IV-VQA (Near OOD)

0.9525AccuracyAdam-SPD(LORA)
0.626980.711490.7960.88051Nov 3, 2024

Evaluation results

MethodLinks
Adam-SPD(LORA)Backbone=PaliGemma-3B, Adaptation Method=LoRA2024.110.9525
Vanilla FT(LORA)Backbone=PaliGemma-3B, Adaptation Method=LoRA2024.110.9443
LP-FT(LORA)Backbone=PaliGemma-3B, Adaptation Method=LoRA2024.110.933
Linear Prob.Backbone=PaliGemma-3B, Adaptation Method=Linear Probing2024.110.8783
WiSE-FT(LORA)Backbone=PaliGemma-3B, Adaptation Method=LoRA2024.110.8506
Zero-ShotBackbone=PaliGemma-3B, Adaptation Method=Zero-Shot2024.110.6395