List your company
ResearchFormerly WizWand

Classification on SVHN VTAB 1k (test)

93.5AccuracyLoRA_mul+VPT_add + PACE
42.5455.776982.23Sep 25, 2024
Updated 8mo ago

Evaluation results

MethodLinks
LoRA_mul+VPT_add + PACEPre-training Method=MAE, Backbone=ViT-B/16, Pre-trained on=ImageNet-1K2024.0993.5
LoRA_mul+VPT_add + PACEPre-training Method=DINO, Backbone=ViT-B/16, Pre-trained on=ImageNet-1K2024.0991.7
FullPre-training Method=MAE, Backbone=ViT-B/16, Pre-trained on=ImageNet-1K2024.0990.1
LoRA_mul+VPT_addPre-training Method=DINO, Backbone=ViT-B/16, Pre-trained on=ImageNet-1K2024.0990
FullPre-training Method=DINO, Backbone=ViT-B/16, Pre-trained on=ImageNet-1K2024.0989.7
LoRA_mul+VPT_addPre-training Method=MAE, Backbone=ViT-B/16, Pre-trained on=ImageNet-1K2024.0989.3
LinearPre-training Method=DINO, Backbone=ViT-B/16, Pre-trained on=ImageNet-1K2024.0950.7
LinearPre-training Method=MAE, Backbone=ViT-B/16, Pre-trained on=ImageNet-1K2024.0944.5