Paimon Voice Clone

Where my speech AI journey started, a VITS voice clone of Paimon from Genshin Impact (2022).

My speech AI journey began as a side quest in the summer of 2022: teaching a model to speak as Paimon from Genshin Impact.

I Used AI to Clone Paimon's Voice (September 2022), 616K views on Bilibili.

Highlights

  • Built and annotated a multi-speaker dataset of ≈48,000 clips (15 hours) from 50 Genshin Impact characters.
  • Engineered the data pipeline with ECAPA-TDNN for speaker classification and Whisper for transcription.
  • Fine-tuned a VITS speech synthesis model on a curated set of Paimon clips.
  • Released the project as a technical demo on Bilibili, where it passed 600,000 views, with a public notebook for anyone to try.

Resources