Image Classification on CIFAR-100 (Accuracy and Calibration Metrics)
93.52AccuracyDEEP ENSEMBLES
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| DEEP ENSEMBLESHeads=122025.10 | 93.52 | 0.09 | 0.22 | 0.01 | 0.01 | |
| MIMOHeads=122025.10 | 92.62 | 0.1 | 0.23 | 0.009 | 0.008 | |
| BATCH ENSEMBLESHeads=122025.10 | 92.19 | 0.11 | 0.26 | 0.008 | 0.006 | |
| SINGLEHeads=122025.10 | 92.15 | 0.11 | 0.25 | 0.007 | 0.005 | |
| LORA ENSEMBLESHeads=122025.10 | 92.14 | 0.11 | 0.26 | 0.007 | 0.006 | |
| HYDRA ENS (CIRC)Heads=10x32025.10 | 92.11 | 0.12 | 0.28 | 0.008 | 0.006 | |
| MC DROPOUTHeads=122025.10 | 92.04 | 0.11 | 0.25 | 0.01 | 0.01 | |
| HYDRA ENS (TAYLOR)Heads=10x32025.10 | 92 | 0.11 | 0.28 | 0.008 | 0.007 | |
| OBA2025.10 | 91.88 | 0.11 | 0.26 | 0.007 | 0.005 | |
| CIRCAVGHeads=102025.10 | 91.67 | 0.12 | 0.3 | 0.01 | 0.01 | |
| TAYLORHeads=102025.10 | 91.59 | 0.12 | 0.31 | 0.01 | 0.01 | |
| PACKED ENSEMBLESHeads=122025.10 | 90.63 | 0.13 | 0.31 | 0.01 | 0.008 |