Image Classification on CIFAR-100 Bucket 1 Retain
90.32AccuracyRetain-only Retain
Evaluation results
| Method | Links | |
|---|---|---|
| Retain-only RetainEvaluation Head=Linear Probe, Backbone=ViT-S/16, Pre-training=ImageNet2026.04 | 90.32 | |
| OriginalEvaluation Head=Output, Backbone=ViT-S/16, Pre-training=ImageNet2026.04 | 90.09 | |
| Retain-only RetainEvaluation Head=Output, Backbone=ViT-S/16, Pre-training=ImageNet2026.04 | 89.95 | |
| OriginalEvaluation Head=Linear Probe, Backbone=ViT-S/16, Pre-training=ImageNet2026.04 | 89.93 | |
| OriginalEvaluation Head=NCC, Backbone=ViT-S/16, Pre-training=ImageNet2026.04 | 89.62 | |
| Retain-only RetainEvaluation Head=NCC, Backbone=ViT-S/16, Pre-training=ImageNet2026.04 | 89.24 | |
| Random-label with CMFEvaluation Head=Output, Backbone=ViT-S/16, Pre-training=ImageNet2026.04 | 86.99 | |
| Random-label with CMFEvaluation Head=NCC, Backbone=ViT-S/16, Pre-training=ImageNet2026.04 | 86.97 | |
| Salun with CMFEvaluation Head=Output, Backbone=ViT-S/16, Pre-training=ImageNet2026.04 | 86.2 | |
| Salun with CMFEvaluation Head=NCC, Backbone=ViT-S/16, Pre-training=ImageNet2026.04 | 86.2 | |
| NegGrad+ with CMFEvaluation Head=NCC, Backbone=ViT-S/16, Pre-training=ImageNet2026.04 | 85.53 | |
| NegGrad+ with CMFEvaluation Head=Output, Backbone=ViT-S/16, Pre-training=ImageNet2026.04 | 85.48 | |
| Random-label with CMFEvaluation Head=Linear Probe, Backbone=ViT-S/16, Pre-training=ImageNet2026.04 | 84.79 | |
| NegGrad+ with CMFEvaluation Head=Linear Probe, Backbone=ViT-S/16, Pre-training=ImageNet2026.04 | 83.47 | |
| Salun with CMFEvaluation Head=Linear Probe, Backbone=ViT-S/16, Pre-training=ImageNet2026.04 | 83.43 |