Efficient Training of Audio Transformers with Patchout

Benchmark Model Rank Results
audio-classification-on-fsd50kPaSST-S#3mAP: 65.55
audio-classification-on-fsd50kPaSST-N-S#6mAP: 64.2
instrument-recognition-on-openmic-2018PaSST#5mean average precision: 0.843
audio-tagging-on-audiosetPaSST#6mean average precision: 0.496
audio-classification-on-esc-50PaSST-S#10Top-1 Accuracy: 96.8PRE-TRAINING DATASET: AudioSet…
audio-classification-on-audiosetPaSST (Ensemble)#12Test mAP: 0.496
audio-classification-on-audiosetPaSST-S (Single)#30Test mAP: 0.471