Target-Speaker Multilingual Speech Generation on TTS multilingual Japanese (test)
3.875WERQwen3-TTS
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen3-TTSTokenizer Frequency=25Hz, Parameters=0.6B, Evaluation Protocol=CustomVoice2026.01 | 3.875 | |
| Qwen3-TTSTokenizer Frequency=25Hz, Parameters=1.7B, Evaluation Protocol=CustomVoice2026.01 | 4.518 | |
| Qwen3-TTSTokenizer Frequency=12Hz, Parameters=1.7B, Evaluation Protocol=CustomVoice2026.01 | 4.924 | |
| GPT-4o-Audio Preview2026.01 | 5.001 | |
| Qwen3-TTSTokenizer Frequency=12Hz, Parameters=0.6B, Evaluation Protocol=CustomVoice2026.01 | 6.877 |