Image Classification on Galaxy 10 DECals
95.42Top-1 AccuracyWaveMix-192/16 (3 level)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| WaveMix-192/16 (3 level)Number of Parameters=28 M, Resolution=256 x 2562022.05 | 95.42 | — | |
| WaveMix-192/16 (level 3)Input Resolution=256 x 256, Pre-trained on ImageNet-1k=true, Augmentation=TrivialAugment [78]2022.05 | 95.42 | — | |
| AstroformerNumber of Parameters=271 M, Resolution=256 x 2562022.05 | 94.86 | — | |
| SwinNumber of Parameters=195 M, Resolution=256 x 2562022.05 | 93.21 | — | |
| Swin V2Number of Parameters=195 M, Resolution=256 x 2562022.05 | 91.23 | — | |
| EfficientNet v2Number of Parameters=117 M, Resolution=256 x 2562022.05 | 90.12 | — | |
| CoAtNetNumber of Parameters=72 M, Resolution=256 x 2562022.05 | 87.45 | — | |
| ViTNumber of Parameters=87 M, Resolution=256 x 2562022.05 | 84.36 | — | |
| Astroformer2022.05 | — | 94.86 | |
| Previous SOTA [4]Pre-train=ImageNet-1k, Parameters=272 M2022.05 | — | 94.86 | |
| WaveMixPre-train=ImageNet-1k, Parameters=28 M2022.05 | — | 95.42 | |
| WaveMixAugmentation=TrivialAugment2022.05 | — | 95.42 |