Show HN: Real-time local TTS (31M params, 5.6x CPU, voice cloning, ONNX)

1 comments

• popalchemist 6 hours ago

  given the architecture, is there a way to force the use of specific phonemes for hard-to-pronounce words? If so that's big