Visual-to-Audio Knowledge Distillation on RAVDESS
74.43AccuracyUCMKD
Evaluation Results
| Method | Links | |
|---|---|---|
| UCMKDBackbone=ResNet-18, Batch size=64, Epoch=100, Optimizer=SGD, Learning rate=1e-2, Unpaired cross-modal setting=true2026.06 | 74.43 | |
| NORMBackbone=ResNet-18, Batch size=64, Epoch=100, Optimizer=SGD, Learning rate=1e-2, Unpaired cross-modal setting=true2026.06 | 70.12 | |
| REVIEWBackbone=ResNet-18, Batch size=64, Epoch=100, Optimizer=SGD, Learning rate=1e-2, Unpaired cross-modal setting=true2026.06 | 68.42 |