Visual Question Answering on CV-VQA Near OOD
69.36AccuracyVanilla FT(LORA)
Evaluation Results
| Method | Links | |
|---|---|---|
| Vanilla FT(LORA)Backbone=PaliGemma-3B, Adaptation Method=LoRA2024.11 | 69.36 | |
| Adam-SPD(LORA)Backbone=PaliGemma-3B, Adaptation Method=LoRA2024.11 | 68.85 | |
| LP-FT(LORA)Backbone=PaliGemma-3B, Adaptation Method=LoRA2024.11 | 65.93 | |
| WiSE-FT(LORA)Backbone=PaliGemma-3B, Adaptation Method=LoRA2024.11 | 64.55 | |
| Linear Prob.Backbone=PaliGemma-3B, Adaptation Method=Linear Probing2024.11 | 63.87 | |
| Zero-ShotBackbone=PaliGemma-3B, Adaptation Method=Zero-Shot2024.11 | 44.72 |