#Agnes Digital (Singing) (Umamusume) (RVC v2) - 375 Epochs

1 messages · Page 1 of 1 (latest)

drifting root
#

-# the closest you'll ever get to my ena shinonome singing model

this character is featured in umamusume (both the anime and game)

trained on 19:25 minutes of her singing data extracted from the game
(cleaned using anvuew's de-reverb, uvr de-echo (in some parts), some eq-ing (for the "meni shuki rush-sshu!" part), eq match, resampled to 40k)

songs used:

  • umapyoi legend
  • make debut!
  • unlimited impact
  • instictive speed
  • girls' legend u
  • we are dreamers!!
  • gaze on me!
  • ms. victoria
  • dramatic journey
  • u.m.a. new world!!
  • starting force
  • meni shuki rush-sshu!

CV: Minori Suzuki

pitch extraction: rmvpe
steps: 28.1k
batch size: 6
pretrain: klm 4 hfg / 40k

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

ezgif-65a3edc3bb2234c0.gif
drifting root
#

for my next model i wanna do a lil experiment and see how well could this pretrain handle small datasets (40 sec) so i'm training her model

#

she spoke less in a movie sooo

drifting root
#

so uhhh about that it doesn't sound really good as i anticipated so i'm not gonna release it

drifting root