# 35-way Speech Classification on Speech Commands 16kHz 35-way (test) Source: [35-way Speech Classification on Speech Commands 16kHz 35-way (test)]() Collection index: [llms.txt]() Data updated: 2026-02-27T09:27:22.942Z Evaluation of 35-way speech classification performance on the test split of the Speech Commands 16kHz 35-way dataset. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | 35-way Speech Classification | | Dataset (including any stated split) | Speech Commands 16kHz 35-way (test) | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 32 of 32 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | MRConv, Fourier | 96.82 | Not recorded | 68fa74801e209131cea61fbf | | MRConv, Fourier | 96.82 | Not recorded | 68fa74801e209131cea61fbf | | Liquid-S4 | 96.78 | Not recorded | 68fa6eab1e209131cea4143a | | Liquid-S4 | 96.78 | Not recorded | 68fa74801e209131cea61fbf | | Liquid-S4 | 96.78 | Not recorded | 68fa74801e209131cea61fbf | | S5 | 96.52 | Not recorded | 68fa6eab1e209131cea4143a | | S5 | 96.52 | Not recorded | 68fa74801e209131cea61fbf | | S5 | 96.52 | Not recorded | 68fa74801e209131cea61fbf | | MRConv, Dilated | 96.47 | Not recorded | 68fa74801e209131cea61fbf | | MRConv, Dilated | 96.47 | Not recorded | 68fa74801e209131cea61fbf | | SGConv | 96.42 | Not recorded | 68fa74801e209131cea61fbf | | SGConv | 96.42 | Not recorded | 68fa74801e209131cea61fbf | | S4-LegS | 96.08 | Not recorded | 68fa6eab1e209131cea4143a | | S4-LegS | 96.08 | Not recorded | 68fa74801e209131cea61fbf | | S4-LegS | 96.08 | Not recorded | 68fa74801e209131cea61fbf | | S4D-LegS | 95.83 | Not recorded | 68fa6eab1e209131cea4143a | | S4D-LegS | 95.83 | Not recorded | 68fa74801e209131cea61fbf | | S4D-LegS | 95.83 | Not recorded | 68fa74801e209131cea61fbf | | ConvNet | 95.51 | Not recorded | 68fa6eab1e209131cea4143a | | ConvNet | 95.51 | Not recorded | 68fa74801e209131cea61fbf | | ConvNet | 95.51 | Not recorded | 68fa74801e209131cea61fbf | | MRConv, Fourier+Sparse | 95.21 | Not recorded | 68fa74801e209131cea61fbf | | MRConv, Fourier+Sparse | 95.21 | Not recorded | 68fa74801e209131cea61fbf | | XResNet-50 | 83.01 | Not recorded | 68fa6eab1e209131cea4143a | | XResNet-50 | 83.01 | Not recorded | 68fa74801e209131cea61fbf | | XResNet-50 | 83.01 | Not recorded | 68fa74801e209131cea61fbf | | ResNet-1 | 77.86 | Not recorded | 68fa6eab1e209131cea4143a | | ResNet-18 | 77.86 | Not recorded | 68fa74801e209131cea61fbf | | ResNet-1 | 77.86 | Not recorded | 68fa74801e209131cea61fbf | | InceptionNet | 61.24 | Not recorded | 68fa6eab1e209131cea4143a | | InceptionNet | 61.24 | Not recorded | 68fa74801e209131cea61fbf | | InceptionNet | 61.24 | Not recorded | 68fa74801e209131cea61fbf | ### MRConv, Fourier Accuracy: 96.82. Result ID: 69a1636f35dfcabb00e6bc08. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Sampling rate":"16 kHz","Input length":"16,000"}. This record does not assert independent reproduction. ### MRConv, Fourier Accuracy: 96.82. Result ID: 69a1637035dfcabb00e6bc30. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Architecture Class":"Convolutional Models","Parameterization":"Fourier"}. This record does not assert independent reproduction. ### Liquid-S4 Accuracy: 96.78. Result ID: 699039443a6c314849ff09a1. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Parameters":"224K","Input length":16000}. This record does not assert independent reproduction. ### Liquid-S4 Accuracy: 96.78. Result ID: 69a1636f35dfcabb00e6bc04. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Sampling rate":"16 kHz","Input length":"16,000"}. This record does not assert independent reproduction. ### Liquid-S4 Accuracy: 96.78. Result ID: 69a1636f35dfcabb00e6bc2c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Architecture Class":"State Space Models"}. This record does not assert independent reproduction. ### S5 Accuracy: 96.52. Result ID: 699039443a6c314849ff09a2. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Parameters":"280K","Input length":16000}. This record does not assert independent reproduction. ### S5 Accuracy: 96.52. Result ID: 69a1636f35dfcabb00e6bc05. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Sampling rate":"16 kHz","Input length":"16,000"}. This record does not assert independent reproduction. ### S5 Accuracy: 96.52. Result ID: 69a1636f35dfcabb00e6bc2d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Architecture Class":"State Space Models"}. This record does not assert independent reproduction. ### MRConv, Dilated Accuracy: 96.47. Result ID: 69a1636f35dfcabb00e6bc07. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Sampling rate":"16 kHz","Input length":"16,000"}. This record does not assert independent reproduction. ### MRConv, Dilated Accuracy: 96.47. Result ID: 69a1637035dfcabb00e6bc2f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Architecture Class":"Convolutional Models","Parameterization":"Dilated"}. This record does not assert independent reproduction. ### SGConv Accuracy: 96.42. Result ID: 69a1636f35dfcabb00e6bc06. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Sampling rate":"16 kHz","Input length":"16,000"}. This record does not assert independent reproduction. ### SGConv Accuracy: 96.42. Result ID: 69a1637035dfcabb00e6bc2e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Architecture Class":"Convolutional Models"}. This record does not assert independent reproduction. ### S4-LegS Accuracy: 96.08. Result ID: 699039443a6c314849ff099f. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Parameters":"307K","Input length":16000}. This record does not assert independent reproduction. ### S4-LegS Accuracy: 96.08. Result ID: 69a1636f35dfcabb00e6bc03. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Sampling rate":"16 kHz","Input length":"16,000"}. This record does not assert independent reproduction. ### S4-LegS Accuracy: 96.08. Result ID: 69a1636f35dfcabb00e6bc2b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Architecture Class":"State Space Models"}. This record does not assert independent reproduction. ### S4D-LegS Accuracy: 95.83. Result ID: 699039443a6c314849ff09a0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Parameters":"306K","Input length":16000}. This record does not assert independent reproduction. ### S4D-LegS Accuracy: 95.83. Result ID: 69a1636f35dfcabb00e6bc02. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Sampling rate":"16 kHz","Input length":"16,000"}. This record does not assert independent reproduction. ### S4D-LegS Accuracy: 95.83. Result ID: 69a1636f35dfcabb00e6bc2a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Architecture Class":"State Space Models"}. This record does not assert independent reproduction. ### ConvNet Accuracy: 95.51. Result ID: 699039443a6c314849ff099e. [Reporting paper]() Recorded evaluation configuration: {"Parameters":"26.2M","Input length":16000}. This record does not assert independent reproduction. ### ConvNet Accuracy: 95.51. Result ID: 69a1636f35dfcabb00e6bc01. [Reporting paper]() Recorded evaluation configuration: {"Sampling rate":"16 kHz","Input length":"16,000"}. This record does not assert independent reproduction. ### ConvNet Accuracy: 95.51. Result ID: 69a1636f35dfcabb00e6bc29. [Reporting paper]() Recorded evaluation configuration: {"Architecture Class":"CNNs"}. This record does not assert independent reproduction. ### MRConv, Fourier+Sparse Accuracy: 95.21. Result ID: 69a1636f35dfcabb00e6bc09. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Sampling rate":"16 kHz","Input length":"16,000"}. This record does not assert independent reproduction. ### MRConv, Fourier+Sparse Accuracy: 95.21. Result ID: 69a1637035dfcabb00e6bc31. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Architecture Class":"Convolutional Models","Parameterization":"Fourier+Sparse"}. This record does not assert independent reproduction. ### XResNet-50 Accuracy: 83.01. Result ID: 699039443a6c314849ff099d. [Reporting paper]() Recorded evaluation configuration: {"Parameters":"904K","Input length":16000}. This record does not assert independent reproduction. ### XResNet-50 Accuracy: 83.01. Result ID: 69a1636f35dfcabb00e6bc00. [Reporting paper]() Recorded evaluation configuration: {"Sampling rate":"16 kHz","Input length":"16,000"}. This record does not assert independent reproduction. ### XResNet-50 Accuracy: 83.01. Result ID: 69a1636f35dfcabb00e6bc28. [Reporting paper]() Recorded evaluation configuration: {"Architecture Class":"CNNs"}. This record does not assert independent reproduction. ### ResNet-1 Accuracy: 77.86. Result ID: 699039443a6c314849ff099c. [Reporting paper]() Recorded evaluation configuration: {"Parameters":"216K","Input length":16000}. This record does not assert independent reproduction. ### ResNet-18 Accuracy: 77.86. Result ID: 69a1636f35dfcabb00e6bbff. [Reporting paper]() Recorded evaluation configuration: {"Sampling rate":"16 kHz","Input length":"16,000"}. This record does not assert independent reproduction. ### ResNet-1 Accuracy: 77.86. Result ID: 69a1636f35dfcabb00e6bc27. [Reporting paper]() Recorded evaluation configuration: {"Architecture Class":"CNNs"}. This record does not assert independent reproduction. ### InceptionNet Accuracy: 61.24. Result ID: 699039443a6c314849ff099b. [Reporting paper]() Recorded evaluation configuration: {"Parameters":"481K","Input length":16000}. This record does not assert independent reproduction. ### InceptionNet Accuracy: 61.24. Result ID: 69a1636f35dfcabb00e6bbfe. [Reporting paper]() Recorded evaluation configuration: {"Sampling rate":"16 kHz","Input length":"16,000"}. This record does not assert independent reproduction. ### InceptionNet Accuracy: 61.24. Result ID: 69a1636f35dfcabb00e6bc26. [Reporting paper]() Recorded evaluation configuration: {"Architecture Class":"CNNs"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.