Speech Command Recognition on Google SC v1 (test)
98.6AccuracySOTA (Topline)
Evaluation Results
| Method | Links | |
|---|---|---|
| SOTA (Topline)Language=En, Number of Classes=122023.03 | 98.6 | |
| BEATsParams=85M, GFLOPs=42.4, RTF=0.2442026.03 | 97.7 | |
| AudioMAEParams=85M, GFLOPs=42.4, RTF=0.2442026.03 | 96.9 | |
| HuBERTParams=94M, GFLOPs=69.6, RTF=0.510, Protocol=frozen pre-trained model2026.03 | 96.3 | |
| wav2vec 2.0Params=94M, GFLOPs=69.6, RTF=0.510, Protocol=frozen pre-trained model2026.03 | 96.2 | |
| SSASTParams=85M, GFLOPs=42.4, RTF=0.2442026.03 | 96 | |
| SpeechPrompt v2 (pGSLM+)Language=En, Backbone=pGSLM, Learnable Verbalizer=true2023.03 | 94.7 | |
| SpeechPrompt v2 (GSLM+)Language=En, Backbone=GSLM, Learnable Verbalizer=true2023.03 | 94.6 | |
| SpeechPrompt v2 (GSLM)Language=En, Backbone=GSLM, Learnable Verbalizer=false2023.03 | 94.5 | |
| SpeechPrompt v2 (pGSLM)Language=En, Backbone=pGSLM, Learnable Verbalizer=false2023.03 | 94.3 | |
| HEARParams=15M, GFLOPs=9.47, RTF=0.0952026.03 | 94.3 |