Fine-tune the pretrained ATST model for sound event detection

Benchmark Model Rank Results
sound-event-detection-on-desedATST-SED#1event-based F1 score: 63.4PSDS1: 0.583PSDS2: 0.810