🎙️**Model**: 💿 **Dataset for curious**:

Dataset for curious:

Robo-Ky [JP] (GUILTY GEAR XX ACCENT CORE PLUS R) — RVC v2 — 1k Epoch

Sample: Робот-д@@@@@б - Красная Плесень Russian song with the robotic voice

For this song I experimented by adding echo effect and combning harvest with transpose = -12 - harvest with transpose = 0 I've found out that this model is not perfect for most songs. Echo effect settings (Audacity): Delay time = 0.1 sec Decay factor = 0.4

На 40 или 48khz тренировал?

48

Not perfect but okay. Harvest with transpose = -3

Update: Hugging Face link updated