README.md

metadata

license: mit
datasets:
  - TLME/Umamusume-voice-transcription

RVC-Umamusume

What is RVC?

https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI

About this repo

Umamusume charcters RVC models.

Each model was trained for 100 epochs with batch size of 12, averaging 40 steps per epoch. On average, each character was trained using 10 to 20 minutes of voice data.

If you think the performance of the model I created is not good, I presented the dataset, and you can train it yourself.

The majority of characters' voices come from :Umamusume-voice-transcription, while a small portion of characters' voices come from: zomehwh/tttttttt .