Image Classification on FGVC and VTAB 24 tasks (test)
78.64AccuracyGPS
Evaluation results
| Method | Links | |
|---|---|---|
| GPSParams (%)=0.36, Model Agnostic=true, No extra Train param.=true, No extra Infer params.=true, Task Adaptive=true, Backbone=ViT-B/162023.12 | 78.64 | |
| SSFParams (%)=0.32, Model Agnostic=false, No extra Train param.=false, No extra Infer params.=true, Task Adaptive=false, Backbone=ViT-B/162023.12 | 76.77 | |
| LoRAParams (%)=0.9, Model Agnostic=false, No extra Train param.=false, No extra Infer params.=true, Task Adaptive=false, Backbone=ViT-B/162023.12 | 75.16 | |
| VPTParams (%)=0.76, Model Agnostic=false, No extra Train param.=false, No extra Infer params.=false, Task Adaptive=false, Backbone=ViT-B/162023.12 | 73.53 | |
| FullParams (%)=100, Model Agnostic=true, No extra Train param.=true, No extra Infer params.=true, Task Adaptive=false, Backbone=ViT-B/162023.12 | 70.36 | |
| BiasParams (%)=0.2, Model Agnostic=true, No extra Train param.=false, No extra Infer params.=true, Task Adaptive=false, Backbone=ViT-B/162023.12 | 67.54 | |
| AdapterParams (%)=0.35, Model Agnostic=false, No extra Train param.=false, No extra Infer params.=false, Task Adaptive=false, Backbone=ViT-B/162023.12 | 60.04 | |
| LinearParams (%)=0.08, Model Agnostic=true, No extra Train param.=true, No extra Infer params.=true, Task Adaptive=false, Backbone=ViT-B/162023.12 | 58.48 |