SSAMBA

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

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

Speech5 results

Keyword SpottingonGoogle Speech Commands V2 35
Accuracy (10-fold)· uses extra data· 2024-05-20
97.4
best: 98.53 (QuaternionNeuralNetwork)
SOTA
SSAMBA: Self-Supervised Audio Representation Learning with Mamba State Space Model arXiv:2405.11831
Keyword SpottingonGoogle Speech Commands
Google Speech Commands V1 12· 2024-05-20
96.9
best: 98.56 (TripletLoss-res15)
SSAMBA: Self-Supervised Audio Representation Learning with Mamba State Space Model arXiv:2405.11831
Keyword SpottingonGoogle Speech Commands
Google Speech Commands V2 35· 2024-05-20
97.4
best: 99.1 (WaveFormer)
SSAMBA: Self-Supervised Audio Representation Learning with Mamba State Space Model arXiv:2405.11831
Speaker IdentificationonVoxCeleb1
Accuracy· uses extra data· 2024-05-20
70.1
best: 96.6 (MSM-MAE)
SSAMBA: Self-Supervised Audio Representation Learning with Mamba State Space Model arXiv:2405.11831
Speaker IdentificationonVoxCeleb1
Top-1 (%)· uses extra data· 2024-05-20
70.1
best: 96.6 (MSM-MAE)
SSAMBA: Self-Supervised Audio Representation Learning with Mamba State Space Model arXiv:2405.11831