Image Classification on CALISA (test)
98.19Clean Accuracyclean data
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| clean dataSampler=DDPM, Backbone=ResNet2026.06 | 98.19 | 2.49 | — | — | |
| one-to-oneSampler=DDIM, Backbone=ResNet2026.06 | 97.88 | 53.79 | 0.18 | 53.09 | |
| all-to-oneSampler=DDIM, Backbone=ResNet2026.06 | 97.72 | 96.5 | 0.02 | 95.8 | |
| clean dataSampler=DDIM, Backbone=ResNet2026.06 | 97.7 | 0.7 | — | — | |
| same-classSampler=DDIM, Backbone=ResNet2026.06 | 97.64 | 28.56 | -0.06 | 27.86 | |
| one-to-oneSampler=DDPM, Backbone=ResNet2026.06 | 97.49 | 100 | -0.7 | 97.51 | |
| all-to-oneSampler=DDPM, Backbone=ResNet2026.06 | 97.46 | 86.9 | -0.73 | 84.41 | |
| same-classSampler=DDPM, Backbone=ResNet2026.06 | 97.28 | 64.36 | -0.91 | 61.87 |