don't mind the image tho lmao
trained on 16 minutes of her speaking (and maybe singing?) from her rabbit hole recording stream (public, but vod members only) (rx11 denoise + de-click)
might sound gnarly on her high range bc it has this 3 second clip thing on a dataset that makes high notes sound gnarly in a model, will possibly retrain it when i'll remove it.
pitch extraction: rmvpe
steps: 15.9k
batch size: 7
pretrain: klm 4.1 / 32k
huggingface: https://huggingface.co/sxndypz/rvc-v2-models/resolve/main/hime_act2.zip
Tags: No tags available
Download Link: https://huggingface.co/sxndypz/rvc-v2-models/resolve/main/hime_act2.zip