Metraipolis › archive › dc_32ee658fe4ea

Paradee: Distilling Kokoro-82M into an 8M-Parameter Single-Voice Text-to-Speech Model — arXiv:2610.06817

Source: arXiv
Original: https://arxiv.org/abs/2610.06817
License (as archived): CC0-1.0
Archived: 2026-10-06T05:27:53.732Z
SHA-256 of the archived text: a786c92c9b2e55871baa7ca08d264c658868e3d1e79dd0a8d822aaa15b5948b9
Stable URL of this page: https://metraipolis.com/documento/dc_32ee658fe4ea

How to cite

Paradee: Distilling Kokoro-82M into an 8M-Parameter Single-Voice Text-to-Speech Model — arXiv:2610.06817 — arXiv, https://arxiv.org/abs/2610.06817 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_32ee658fe4ea (Metraipolis archive dc_32ee658fe4ea, sha256 a786c92c9b2e55871baa7ca08d264c658868e3d1e79dd0a8d822aaa15b5948b9)

BibTeX

@misc{dc32ee658fe4ea,
  title = {Paradee: Distilling Kokoro-82M into an 8M-Parameter Single-Voice Text-to-Speech Model — arXiv:2610.06817},
  howpublished = {arXiv},
  url = {https://arxiv.org/abs/2610.06817},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_32ee658fe4ea; sha256 a786c92c9b2e55871baa7ca08d264c658868e3d1e79dd0a8d822aaa15b5948b9},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.06817. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

We distill Kokoro-82M, a widely used open text-to-speech model with 54 voices, into Paradee, an 8.07M-parameter model that speaks one of them. Paradee keeps Kokoro's architecture with much narrower layers, and each of its two halves is trained separately against the frozen teacher. It has 10x fewer parameters and needs 15x less compute. We first synthesize a corpus with the teacher and keep its durations, pitch, energy and phoneme features. We then train a small text side to predict these values, and a small decoder to turn the teacher's saved values into the teacher's audio, first with spectral losses and then adversarially. Finally, we connect the two halves and quantize the weights to int8. It needs no alignment learning and no joint training, and it runs on one laptop. Stored in int8, Paradee is 8.5 MB, runs 25x faster than real time on one CPU thread, and scores 4.41 on UTMOS against the teacher's 4.52. The student initially kept a slight buzz, which we trace to the phase of voiced speech between 2 and 8 kHz. A phase-locking filter applied after synthesis removes most of it, with no training and no extra parameters. Code, model files and audio samples are at this https URL

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Sound (cs.SD) ; Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG); Audio and Speech Processing (eess.AS).
Pagina del lavoro: https://arxiv.org/abs/2610.06817 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-06T05:24:06.194Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).