List your company
ResearchFormerly WizWand

Language Identification on Low-resource Dialectal Speech 5 languages (test)

98.06Detection AccuracyCLD
47.8860.907573.93586.9625May 22, 2026
Updated 5mo ago

Evaluation results

MethodLinks
CLDASR Model Backbone=Whisper-Large-v3 (WSP-L)2026.0598.06
CLDASR Model Backbone=Whisper-Small (WSP)2026.0597.15
CLDASR Model Backbone=MMS-1B2026.0597.02
NNASR Model Backbone=Whisper-Large-v3 (WSP-L)2026.0596.05
Kernel SVMASR Model Backbone=Whisper-Large-v3 (WSP-L)2026.0595.82
Linear SVMASR Model Backbone=Whisper-Large-v3 (WSP-L)2026.0595.01
Kernel SVMASR Model Backbone=Whisper-Small (WSP)2026.0594.31
Linear SVMASR Model Backbone=Whisper-Small (WSP)2026.0593.92
NNASR Model Backbone=MMS-1B2026.0586.12
DefaultASR Model Backbone=Whisper-Large-v3 (WSP-L)2026.0580.33
NNASR Model Backbone=Whisper-Small (WSP)2026.0577.37
DefaultASR Model Backbone=Whisper-Small (WSP)2026.0571.54
KNNASR Model Backbone=Whisper-Large-v3 (WSP-L)2026.0571.45
DefaultASR Model Backbone=MMS-1B2026.0567.01
KNNASR Model Backbone=Whisper-Small (WSP)2026.0561.23
Kernel SVMASR Model Backbone=MMS-1B2026.0557.01
Linear SVMASR Model Backbone=MMS-1B2026.0556.53
KNNASR Model Backbone=MMS-1B2026.0549.81