Text-to-Speech on LibriSpeech-PC (test)
2.02WERWavTTS
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| WavTTSEvaluation Protocol=zero-shot2026.06 | 2.02 | 4.36 | — | — | — | |
| FlexiSLMFR (Hz)=12.52026.06 | 2.14 | — | — | — | — | |
| Ground Truth2026.06 | 2.23 | 4.1 | — | — | — | |
| VITSLJEvaluation Protocol=supervised2026.06 | 2.23 | 4.36 | — | — | — | |
| CosyVoice2Type=Only TTS Models, Params=0.6B, Data-hrs=167K2025.10 | 2.23 | 4.38 | 0.66 | 3.98 | 4.29 | |
| Ground TruthType=Ground Truth2025.10 | 2.43 | 4.07 | 0.69 | 3.82 | 4.33 | |
| FlexiSLMFR (Hz)=8.02026.06 | 2.47 | — | — | — | — | |
| MaskGCTType=Only TTS Models, Params=1.0B, Data-hrs=100K2025.10 | 2.49 | 3.85 | 0.66 | 3.92 | 4.12 | |
| F5-TTSType=Only TTS Models, Params=0.3B, Data-hrs=100K2025.10 | 2.54 | 3.84 | 0.66 | 3.9 | 3.98 | |
| FireRedTTSType=Only TTS Models, Params=0.6B, Data-hrs=248K2025.10 | 2.69 | 3.91 | 0.47 | 3.85 | 4.01 | |
| FlexiSLMFR (Hz)=6.252026.06 | 2.87 | — | — | — | — | |
| Spark-TTSType=Only TTS Models, Params=0.3B, Data-hrs=102K2025.10 | 2.89 | 4.23 | 0.55 | 3.95 | 4.11 | |
| JETSEvaluation Protocol=supervised2026.06 | 3 | 4.34 | — | — | — | |
| Qwen2.5-OmniFR (Hz)=252026.06 | 3.18 | — | — | — | — | |
| CosyVoiceFR (Hz)=252026.06 | 3.2 | — | — | — | — | |
| Qwen3-OmniFR (Hz)=12.52026.06 | 3.34 | — | — | — | — | |
| CosyVoiceType=Only TTS Models, Params=0.4B, Data-hrs=170K2025.10 | 3.59 | 4.17 | 0.66 | 3.96 | 4.15 | |
| UniVoiceType=Unified Models, Params=0.4B, Data-hrs=50K2025.10 | 4.06 | 3.72 | 0.56 | 3.88 | 3.96 | |
| FlexiSLMFR (Hz)=5.02026.06 | 4.16 | — | — | — | — | |
| OpusLM-7BType=Unified Models, Params=7B, Data-hrs=213K2025.10 | 4.6 | — | — | — | — | |
| UniVoice-TTSType=Only TTS Models, Params=0.4B, Data-hrs=50K2025.10 | 4.66 | 3.92 | 0.56 | 3.86 | 4.06 | |
| SpeechT5Type=Unified Models, Params=0.14B, Data-hrs=0.96K2025.10 | 5.91 | 3.32 | 0.33 | 3.35 | 3.36 | |
| LauraGPTType=Unified Models, Params=2.0B, Data-hrs=60K2025.10 | 8.62 | — | — | — | — | |
| VITSVCTKEvaluation Protocol=supervised2026.06 | 10.23 | 4.03 | — | — | — | |
| Mimo-AudioFR (Hz)=252026.06 | 14.72 | — | — | — | — | |
| OpusLM-0.4BType=Unified Models, Params=0.4B, Data-hrs=213K2025.10 | 19.8 | — | — | — | — | |
| WaveGrad 2Evaluation Protocol=supervised2026.06 | 33.77 | 3.05 | — | — | — |