-# this model is a bit scuffed but not as scuffed as the version trained on a dataset including meni shuki rush-sshu!
trained on 24 minutes of her singing data extracted from the game
(cleaned using anvuew's de-reverb, uvr de-echo (in some parts), anvuew karaoke (specifically in the "soshite minna no" part + bs splitter for the screaming part at the end), resampled to 40k)
songs used:
-
umapyoi legend
-
never looking back
-
make debut!
-
winning the soul
-
girls' legend u
-
run for our dream
-
we are dreamers!!
-
gaze on me
-
ms. victoria
-
everlasting beats
-
dramatic journey
-
l'arc de gloire
-
tracen ondo
-
soshire minna no
-
u.m.a. new world!!
-
starting force
CV: Hina Tachibana
pitch extraction: rmvpe
steps: 24.1k
batch size: 6
pretrain: klm 4 hfg / 40k
hf: https://huggingface.co/sxndypz/rvc-v2-models/resolve/main/satonodiamond_sing.zip