1 repository on SrcLog
Add it to a GitHub README or profile β the card updates as the ranking changes:
Synthesize 32 kHz reference-conditioned speech with a two-stage TTS model delivering 83.28 speaker similarity on Seed-TTS-Eval.