Automatic Drum Transcription on E-GMD (test)
89.68Onset F1N2N
Evaluation results
| Method | Links | ||
|---|---|---|---|
| N2NSampling steps=10, Training data=E-GMD, Model Type=Generative, Architecture=Transformer, Velocity Capability=Yes2025.09 | 89.68 | 82.8 | |
| N2NSampling steps=5, Training data=E-GMD, Model Type=Generative, Architecture=Transformer, Velocity Capability=Yes2025.09 | 89.24 | 82.56 | |
| N2NSampling steps=1, Training data=E-GMD, Model Type=Generative, Architecture=Transformer, Velocity Capability=Yes2025.09 | 88 | 79.1 | |
| hFT-TransformerTraining data=E-GMD, Model Type=Discriminative, Architecture=Transformer, Velocity Capability=Yes2025.09 | 86.31 | 80.16 | |
| OaF DrumsTraining data=E-GMD, Model Type=Discriminative, Architecture=CRNN, Velocity Capability=Yes2025.09 | 83.4 | 61.7 | |
| DT-EnsembleTraining data=TMIDT(-Bal), MDB, ENST, RBMA, Model Type=Discriminative, Architecture=CRNN, Velocity Capability=No2025.09 | 64.98 | — | |
| ADTOFTraining data=ADTOF, Model Type=Discriminative, Architecture=CRNN, Velocity Capability=No2025.09 | 44.95 | — |