Image Classification on Balanced IFCB Plankton Dataset (test)
0.9508AccuracyEfficientNet-B3
Evaluation results
| Method | Links | ||
|---|---|---|---|
| EfficientNet-B3Evaluation Protocol=Linear probe, Params=11M2026.01 | 0.9508 | 0.9447 | |
| EfficientNet-B0Evaluation Protocol=Linear probe, Params=4M2026.01 | 0.949 | 0.9455 | |
| EfficientNet-V2-SEvaluation Protocol=Linear probe, Params=22.5M2026.01 | 0.9302 | 0.9257 | |
| ResNet-50Evaluation Protocol=Linear probe, Params=25.5M2026.01 | 0.9281 | 0.9229 | |
| Our DiffusionEvaluation Protocol=Linear probe, Params=57M2026.01 | 0.924 | 0.9114 | |
| ViT-B/16Evaluation Protocol=Linear probe, Params=87M2026.01 | 0.9209 | 0.9128 | |
| Masked Autoencodersbackbone=ViT-B/16, Evaluation Protocol=Linear probe, Params=87M2026.01 | 0.8467 | 0.8214 | |
| DINOv3backbone=ViT-B/16, Evaluation Protocol=Linear probe, Params=87M2026.01 | 0.8031 | 0.6145 | |
| MLP on featuresEvaluation Protocol=Linear probe, Params=0.1M2026.01 | 0.7602 | 0.7206 |