SSAST-PATCH

Reported on 5 benchmarks across 4 tasks · 1 paper · 2 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Speech3 results

Speaker IdentificationonVoxCeleb1
Accuracy· uses extra data· 2021-10-19
64.2
best: 96.6 (MSM-MAE)
SSAST: Self-Supervised Audio Spectrogram Transformer arXiv:2110.09784
Speaker IdentificationonVoxCeleb1
Top-1 (%)· uses extra data· 2021-10-19
64.2
best: 96.6 (MSM-MAE)
SSAST: Self-Supervised Audio Spectrogram Transformer arXiv:2110.09784
Spoken Command RecognitiononSpeech Command v2
Accuracy· 2021-10-19
98
best: 98.1 (SSAST-FRAME)
SSAST: Self-Supervised Audio Spectrogram Transformer arXiv:2110.09784

Audio1 result

Audio ClassificationonBalanced Audio Set
Mean AP· 2021-10-19
31
best: 42.4 (EquiAV)
SOTA
SSAST: Self-Supervised Audio Spectrogram Transformer arXiv:2110.09784

Methodology1 result

ClassificationonBalanced Audio Set
Mean AP· 2021-10-19
31
best: 42.4 (EquiAV)
SOTA
SSAST: Self-Supervised Audio Spectrogram Transformer arXiv:2110.09784