Image Classification on CIFAR-100 LT (500:1 ratio) (test)
3.938LossIGLU-Approx
Evaluation results
| Method | Links | ||
|---|---|---|---|
| IGLU-Approxsigma=0.1, Backbone=ResNet-202026.03 | 3.938 | 20.6 | |
| IGLUsigma=0.1, Backbone=ResNet-202026.03 | 3.957 | 17.99 | |
| IGLU-Approxsigma=0.5, Backbone=ResNet-202026.03 | 5.454 | 28.84 | |
| IGLUsigma=0.5, Backbone=ResNet-202026.03 | 5.659 | 28.94 | |
| ReLUBackbone=ResNet-202026.03 | 5.711 | 26.66 | |
| IGLUsigma=10, Backbone=ResNet-202026.03 | 6.014 | 28.37 | |
| IGLU-Approxsigma=1, Backbone=ResNet-202026.03 | 6.089 | 28.13 | |
| MishBackbone=ResNet-202026.03 | 6.132 | 28.79 | |
| IGLUsigma=5, Backbone=ResNet-202026.03 | 6.182 | 28.8 | |
| IGLUsigma=1, Backbone=ResNet-202026.03 | 6.192 | 29.03 | |
| IGLU-Approxsigma=10, Backbone=ResNet-202026.03 | 6.211 | 27.54 | |
| GELUBackbone=ResNet-202026.03 | 6.237 | 26.77 | |
| HardswishBackbone=ResNet-202026.03 | 6.251 | 26.15 | |
| SiLUBackbone=ResNet-202026.03 | 6.377 | 28.18 | |
| IGLU-Approxsigma=5, Backbone=ResNet-202026.03 | 6.381 | 28.38 |