Image Classification on DTD (Testing Accuracy)
54.68Testing AccuracyGrad-Mimic
Evaluation results
| Method | Links | |
|---|---|---|
| Grad-MimicNoise Level=0.4, Evaluation Protocol=linear probing, Backbone=ViT-B/162025.01 | 54.68 | |
| Rho-LossNoise Level=0.4, Evaluation Protocol=linear probing, Backbone=ViT-B/162025.01 | 54.1 | |
| Grad-MimicNoise Level=0.5, Evaluation Protocol=linear probing, Backbone=ViT-B/162025.01 | 54.1 | |
| Rho-LossNoise Level=0.5, Evaluation Protocol=linear probing, Backbone=ViT-B/162025.01 | 52.39 | |
| SGDNoise Level=0.4, Evaluation Protocol=linear probing, Backbone=ViT-B/162025.01 | 51.91 | |
| Grad-MatchNoise Level=0.4, Evaluation Protocol=linear probing, Backbone=ViT-B/162025.01 | 51.81 | |
| Grad-MimicNoise Level=0.6, Evaluation Protocol=linear probing, Backbone=ViT-B/162025.01 | 50.43 | |
| Rho-LossNoise Level=0.6, Evaluation Protocol=linear probing, Backbone=ViT-B/162025.01 | 48.46 | |
| SGDNoise Level=0.5, Evaluation Protocol=linear probing, Backbone=ViT-B/162025.01 | 47.71 | |
| Grad-MatchNoise Level=0.5, Evaluation Protocol=linear probing, Backbone=ViT-B/162025.01 | 47.55 | |
| AGRANoise Level=0.4, Evaluation Protocol=linear probing, Backbone=ViT-B/162025.01 | 41.81 | |
| SGDNoise Level=0.6, Evaluation Protocol=linear probing, Backbone=ViT-B/162025.01 | 41.44 | |
| Grad-MatchNoise Level=0.6, Evaluation Protocol=linear probing, Backbone=ViT-B/162025.01 | 41.33 | |
| AGRANoise Level=0.5, Evaluation Protocol=linear probing, Backbone=ViT-B/162025.01 | 36.28 | |
| Grad-NormNoise Level=0.4, Evaluation Protocol=linear probing, Backbone=ViT-B/162025.01 | 36.22 | |
| AGRANoise Level=0.6, Evaluation Protocol=linear probing, Backbone=ViT-B/162025.01 | 31.49 | |
| Grad-NormNoise Level=0.5, Evaluation Protocol=linear probing, Backbone=ViT-B/162025.01 | 30.59 | |
| Grad-NormNoise Level=0.6, Evaluation Protocol=linear probing, Backbone=ViT-B/162025.01 | 25.32 |