Visual-to-Audio Knowledge Distillation on AVE
52.24AccuracyUCMKD
Evaluation results
| Method | Links | |
|---|---|---|
| UCMKDBackbone=ResNet-18, Batch size=64, Epoch=100, Optimizer=SGD, Learning rate=1e-2, Unpaired cross-modal setting=true2026.06 | 52.24 | |
| NORMBackbone=ResNet-18, Batch size=64, Epoch=100, Optimizer=SGD, Learning rate=1e-2, Unpaired cross-modal setting=true2026.06 | 50.49 | |
| REVIEWBackbone=ResNet-18, Batch size=64, Epoch=100, Optimizer=SGD, Learning rate=1e-2, Unpaired cross-modal setting=true2026.06 | 46.27 |