LiFT

Reported on 3 benchmarks across 1 task · 1 paper · 1 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Methodology3 results

Representation LearningonImageNet
Average Drop· 2024-03-21
66.9
SOTA
LiFT: A Surprisingly Simple Lightweight Feature Transform for Dense ViT Descriptors arXiv:2403.14625
Representation LearningonImageNet
ADCC· 2024-03-21
53
best: 73.3 (JAFAR)
LiFT: A Surprisingly Simple Lightweight Feature Transform for Dense ViT Descriptors arXiv:2403.14625
Representation LearningonImageNet
Average Increase· 2024-03-21
8.7
best: 32.7 (SAPA)
LiFT: A Surprisingly Simple Lightweight Feature Transform for Dense ViT Descriptors arXiv:2403.14625