Acoustic Scene Classification on FSD50K (test)
54.1mAPLarge Transformer with multi-scale filters
Evaluation Results
| Method | Links | |
|---|---|---|
| Large Transformer with multi-scale filters# Param=2.3M2021.05 | 54.1 | |
| Large 6-Layer Transformer with Pooling# Param=2.3M2021.05 | 53.7 | |
| Large 6-Layer Transformer# Param=2.3M2021.05 | 52.5 | |
| Small Transformer# Param=0.9M2021.05 | 46.9 | |
| VGG-like# Param=0.27M2021.05 | 43.4 | |
| DenseNet-121# Param=12.5M2021.05 | 42.5 | |
| CRNN# Param=0.96M2021.05 | 41.7 | |
| ResNet-18# Param=11.3M2021.05 | 37.3 |