Zero-Shot Learning on CUB (Proposed Split)
70.9Top-1 Accuracy (Class-Avg)W/O ADA (Ours)
Evaluation results
| Method | Links | |
|---|---|---|
| W/O ADA (Ours)Backbone=ResNet-101, Evaluation Protocol=Inductive, Split=Proposed Split (PS)2019.06 | 70.9 | |
| SYNCBackbone=ResNet-101, Evaluation Protocol=Inductive, Split=Proposed Split (PS)2019.06 | 55.6 | |
| SJEBackbone=ResNet-101, Evaluation Protocol=Inductive, Split=Proposed Split (PS)2019.06 | 53.9 | |
| ESZSLBackbone=ResNet-101, Evaluation Protocol=Inductive, Split=Proposed Split (PS)2019.06 | 53.9 | |
| CVAE-ZSLBackbone=ResNet-101, Evaluation Protocol=Inductive, Split=Proposed Split (PS)2019.06 | 52.1 | |
| DEVISEBackbone=ResNet-101, Evaluation Protocol=Inductive, Split=Proposed Split (PS)2019.06 | 52 | |
| DEMBackbone=ResNet-101, Evaluation Protocol=Inductive, Split=Proposed Split (PS)2019.06 | 51.7 | |
| LATEMBackbone=ResNet-101, Evaluation Protocol=Inductive, Split=Proposed Split (PS)2019.06 | 49.3 | |
| GFZSLBackbone=ResNet-101, Evaluation Protocol=Inductive, Split=Proposed Split (PS)2019.06 | 49.2 | |
| SSEBackbone=ResNet-101, Evaluation Protocol=Inductive, Split=Proposed Split (PS)2019.06 | 43.9 | |
| DAPBackbone=ResNet-101, Evaluation Protocol=Inductive, Split=Proposed Split (PS)2019.06 | 40 | |
| CMTBackbone=ResNet-101, Evaluation Protocol=Inductive, Split=Proposed Split (PS)2019.06 | 34.6 | |
| CONSEBackbone=ResNet-101, Evaluation Protocol=Inductive, Split=Proposed Split (PS)2019.06 | 34.3 | |
| SAEBackbone=ResNet-101, Evaluation Protocol=Inductive, Split=Proposed Split (PS)2019.06 | 33.3 | |
| IAPBackbone=ResNet-101, Evaluation Protocol=Inductive, Split=Proposed Split (PS)2019.06 | 24 |