Sound Event Separation on FSD-Kaggle 3 Sound 2018
9.27SI-SDRTF-Locoformer
Evaluation results
| Method | Links | |
|---|---|---|
| TF-LocoformerLearning Paradigm=Supervised2025.12 | 9.27 | |
| UnasDiffLearning Paradigm=Unsupervised, Strategy=Gradient+Initialization2025.12 | 5.15 | |
| Conv-TasNetLearning Paradigm=Supervised2025.12 | 5.02 | |
| DSG’s σ(t)-prop. γ(t)Learning Paradigm=Unsupervised, Strategy=Gradient+Initialization2025.12 | 4.75 | |
| DPS’s constant γ(t)Learning Paradigm=Unsupervised, Strategy=Gradient+Initialization2025.12 | 0.12 | |
| Analytical SamplingLearning Paradigm=Unsupervised, Strategy=Analytical Sampling2025.12 | -2.18 | |
| Unprocessed2025.12 | -3.13 |