-# a remaster of the model i made 2 years ago

my first model trained on og pretrain after a while.

trained on (almost) 21 minutes of her speaking from the try not to laugh stream
(cleaned using hyperace v2 voc, mel denoise, anvuew room de-reverb, anvuew karaoke (in some parts), some eq, resampled to 32k)
-
pitch extraction: rmvpe
steps: 20.1k
batch size: 6
pretrain: original / 32k

hf: https://huggingface.co/sxndypz/rvc-v2-models/resolve/main/roca_act2.zip

Additional Details:

Tags: No tags available

Download Link: https://huggingface.co/sxndypz/rvc-v2-models/resolve/main/roca_act2.zip Hosted on Hugging Face