Image Classification on TinyImgNet
66.08AccuracyTeacher
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| TeacherBackbone=ResNet182026.04 | 66.08 | — | — | |
| VLM BaselineIR=1, Backbone=VLM2026.01 | 44.3 | — | — | |
| DFHL-RSBackbone=ResNet18, Synthetic Dataset Size=500K2026.04 | 30.52 | — | — | |
| DIP-KDBackbone=ResNet18, Synthetic Dataset Size=500K2026.04 | 28.03 | — | — | |
| RN50 BaselineIR=1, Backbone=ResNet-502026.01 | 25.7 | — | — | |
| IDEALBackbone=ResNet18, Synthetic Dataset Size=500K2026.04 | 23.76 | — | — | |
| ZSDB3Backbone=ResNet18, Synthetic Dataset Size=500K2026.04 | 3.29 | — | — | |
| NaiveKDBackbone=ResNet18, Synthetic Dataset Size=500K2026.04 | 0.75 | — | — | |
| PRISMIR=1, Attack Type=Blend, Backbone=ResNet-502026.01 | — | 17.8 | 0 | |
| PRISMIR=1, Attack Type=BPP, Backbone=ResNet-502026.01 | — | 16.3 | 0 | |
| PRISMIR=1, Attack Type=CTRL, Backbone=ResNet-502026.01 | — | 19.6 | 0 |