Representation Learning on ImageNet
Metric: Average Drop (higher is better)
Metric: Average Drop (higher is better)
| # | Model↕ | Average Drop▼ | Augmentations | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | LiFT | 66.9 | No | LiFT: A Surprisingly Simple Lightweight Feature ... | 2024-03-21 | Code |
| 2 | CARAFE | 49.9 | No | CARAFE: Content-Aware ReAssembly of FEatures | 2019-05-06 | Code |
| 3 | Bilinear | 19 | No | - | - | - |
| 4 | Dysample | 17.8 | No | Learning to Upsample by Learning to Sample | 2023-08-29 | Code |
| 5 | JAFAR | 17.4 | No | JAFAR: Jack up Any Feature at Any Resolution | 2025-06-10 | Code |
| 6 | FeatUp (JBU) | 15.3 | No | FeatUp: A Model-Agnostic Framework for Features ... | 2024-03-15 | Code |
| 7 | ReSFU | 14.5 | No | A Refreshed Similarity-based Upsampler for Direc... | 2024-07-02 | Code |
| 8 | SAPA | 8.5 | No | SAPA: Similarity-Aware Point Affiliation for Fea... | 2022-09-26 | Code |