Image Classification on DTD (Accuracy, NLL, ECE, Brier)
82.98AccuracySOAP
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| SOAPOptimizer=SOAP, Monte Carlo samples (BMA)=1, Backbone=ViT-B/16, Pre-training=OpenAI CLIP, Fine-tuning protocol=linear layers (99% of parameters)2026.06 | 82.98 | 0.696 | 7.04 | 0.263 | |
| EVONOptimizer=EVON, Monte Carlo samples (BMA)=1, Backbone=ViT-B/16, Pre-training=OpenAI CLIP, Fine-tuning protocol=linear layers (99% of parameters)2026.06 | 82.15 | 0.787 | 8.86 | 0.281 | |
| EVONOptimizer=EVON, Monte Carlo samples (BMA)=32, Backbone=ViT-B/16, Pre-training=OpenAI CLIP, Fine-tuning protocol=linear layers (99% of parameters)2026.06 | 81.99 | 0.633 | 3.53 | 0.26 | |
| AdamWOptimizer=AdamW, Monte Carlo samples (BMA)=1, Backbone=ViT-B/16, Pre-training=OpenAI CLIP, Fine-tuning protocol=linear layers (99% of parameters)2026.06 | 81.2 | 0.813 | 9.21 | 0.292 | |
| IVONOptimizer=IVON, Monte Carlo samples (BMA)=32, Backbone=ViT-B/16, Pre-training=OpenAI CLIP, Fine-tuning protocol=linear layers (99% of parameters)2026.06 | 80.6 | 0.67 | 2.47 | 0.282 | |
| IVONOptimizer=IVON, Monte Carlo samples (BMA)=1, Backbone=ViT-B/16, Pre-training=OpenAI CLIP, Fine-tuning protocol=linear layers (99% of parameters)2026.06 | 80.48 | 0.731 | 4.93 | 0.291 |