Image Classification on ImageNet-1k 1.0 (val) (Top-1/Top-5 Accuracy, CE Loss)
59.324Top-1 AccuracyGCond
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| GCondBackbone=ConvNeXtV2-Base2025.09 | 59.324 | 81.88 | 1.8793 | |
| CAGradBackbone=ConvNeXtV2-Base2025.09 | 56.02 | 78.57 | 2.0155 | |
| PCGradBackbone=ConvNeXtV2-Base2025.09 | 55.304 | 78.084 | 2.034 | |
| BaselineBackbone=ConvNeXtV2-Base2025.09 | 55.08 | 78.058 | 2.0682 | |
| GradNormBackbone=ConvNeXtV2-Base2025.09 | 54.874 | 78.008 | 2.0577 | |
| GradDropBackbone=ConvNeXtV2-Base2025.09 | 54.668 | 77.782 | 2.0735 |