Image Classification on CIFAR-100-LT (100:1 ratio) (test)
3.426LossIGLU
Evaluation results
| Method | Links | ||
|---|---|---|---|
| IGLUsigma=0.1, Backbone=ResNet-202026.03 | 3.426 | 23.05 | |
| IGLU-Approxsigma=0.1, Backbone=ResNet-202026.03 | 3.452 | 23.01 | |
| IGLU-Approxsigma=0.5, Backbone=ResNet-202026.03 | 4.784 | 32.46 | |
| IGLUsigma=0.5, Backbone=ResNet-202026.03 | 4.878 | 31.45 | |
| IGLU-Approxsigma=1, Backbone=ResNet-202026.03 | 5.154 | 32.43 | |
| IGLU-Approxsigma=10, Backbone=ResNet-202026.03 | 5.347 | 30.52 | |
| IGLUsigma=10, Backbone=ResNet-202026.03 | 5.35 | 30.89 | |
| IGLUsigma=1, Backbone=ResNet-202026.03 | 5.352 | 31.64 | |
| ReLUBackbone=ResNet-202026.03 | 5.436 | 30.01 | |
| IGLU-Approxsigma=5, Backbone=ResNet-202026.03 | 5.447 | 31.74 | |
| IGLUsigma=5, Backbone=ResNet-202026.03 | 5.456 | 30.15 | |
| MishBackbone=ResNet-202026.03 | 5.578 | 31.83 | |
| GELUBackbone=ResNet-202026.03 | 5.603 | 30.17 | |
| HardswishBackbone=ResNet-202026.03 | 5.685 | 30.08 | |
| SiLUBackbone=ResNet-202026.03 | 5.693 | 31.31 |