List your company

Text-to-Speech on LibriSpeech-PC subset B

1.95WERCosyVoice3
1.88442.32722.773.2128Jun 8, 2026

Evaluation results

MethodLinks
CosyVoice3Type=AR, Params.=0.5B, Dataset=1000k Multi.2026.061.9571.84.28
CosyVoice2Type=AR, Params.=0.5B, Dataset=170k Multi.2026.062.0565.74.38
TLDRType=Ours, k=4, Params.=0.5B + 136.2M, Dataset=0.6k LibriTTS2026.062.15714.24
TLDRType=Ours, k=6, Params.=0.5B + 136.2M, Dataset=0.6k LibriTTS2026.062.270.84.23
Ground-truthParams.=-, Dataset=-2026.062.23694.1
DiTARType=AR, Params.=0.6B, Dataset=100k Emilia2026.062.39674.22
F5-TTSType=NAR, NFE=32, Params.=0.3B, Dataset=100k Emilia2026.062.42663.88
TLDRType=Ours, k=8, Params.=0.5B + 136.2M, Dataset=0.6k LibriTTS2026.062.5370.94.23
FireRedTTSType=AR, Params.=0.6B, Dataset=248k Multi.2026.062.6947—
MaskGCTType=NAR, NFE=50, Params.=1.1B, Dataset=100k Emilia2026.062.72693.9
E2 TTSType=NAR, NFE=32, Params.=0.3B, Dataset=100k Emilia2026.062.95693.56
CosyVoiceType=AR, Params.=0.3B, Dataset=170k Multi.2026.063.59664.14