Image Classification on ImageNet (test) (Standard and Robust Accuracy)
76.23Standard AccuracyNADD
Evaluation results
| Method | Links | ||
|---|---|---|---|
| NADDDefense Type=AP, Attack setting=L-infinity norm (epsilon = 4/255), Attack=AutoAttack2026.01 | 76.23 | 44.23 | |
| Nie et al.Defense Type=AP, Attack setting=L-infinity norm (epsilon = 4/255), Attack=AutoAttack2026.01 | 71.48 | 38.71 | |
| Lee et al.Defense Type=AP, Attack setting=L-infinity norm (epsilon = 4/255), Attack=AutoAttack, Number of diffusion cycles=82026.01 | 70.74 | 42.15 | |
| Bai et al.Defense Type=AT, Attack setting=L-infinity norm (epsilon = 4/255), Attack=AutoAttack2026.01 | 67.38 | 35.51 | |
| Salman et al.Defense Type=AT, Attack setting=L-infinity norm (epsilon = 4/255), Attack=AutoAttack2026.01 | 63.86 | 39.11 | |
| Engstrom et al.Defense Type=AT, Attack setting=L-infinity norm (epsilon = 4/255), Attack=AutoAttack2026.01 | 62.42 | 33.2 | |
| Wong et al.Defense Type=AT, Attack setting=L-infinity norm (epsilon = 4/255), Attack=AutoAttack2026.01 | 53.83 | 28.04 |