https://uberduck.ai/ logo
Join Discord
Powered by
# machine-learning
  • a
    What character?
  • z
    Filburt
  • a
    Ahhh
  • z
    I may be bad at collecting wavs... or not
  • u
    been training a Steve Burns (blue's clues) model since yesterday. Only at 6 minutes of audio rn so it's not so great (has no idea how to pronounce certain words still) but it's progress i guess.
  • w
    sheriff woody.
  • h
    Aight
  • h
    Any typical ETA for a voice request?
  • h
    Like how long would it take to be trained and uploaded?
  • p
    no timeline
  • p
    if you want it quickly you need to do it yourself
  • h
    The voice set's already provided though so I guess part of the equation has been done
  • h
    yeah
  • p
    which one you mean?
  • h
    Snow White
  • p
    is it in #843619158834020393 ?
  • h
  • p
    i'll take that as a no
  • h
    wdym?
  • p
    read again what I wrote
  • p
    also there are no transcripts
  • h
    yeah it's more of a voice request but I've provided a link to the voice clips to be used to save you guys the trouble
  • h
    unlike most voice requests which simply ask to have X done but there's next to no links to where can you get clean clips
  • p
    yes, ok; but unless you provide a ready to train dataset you shouldn't hold your breath for the voice to be added
  • h
    I'm not in a rush though so it's OK
  • s
    sounds good
  • u
    This is my first working model (even if I know that someone already made one), trained for 1 hour with only 2 minutes of audio. Isn't it too bad?
  • h
    Sounds decent enough though a bit incoherent
  • h
    or maybe my ear's just being an arse
  • u
    Yeah some syllables are missing
1...128129130...1068Latest