https://uberduck.ai/ logo
Join Discord
Powered by
# machine-learning
  • a
    Usually 20-30 minutes. Some have been able to make it work with much less audio.
  • u
    How long does that take to transcribe?
  • h
    Howdy
  • u
    dyhow'
  • h
    Just came here to ask for help on how to use a raw voice set I have of Snow White
  • u
    😳✋ hello-
  • h
    27 minutes of audio from what I've gathered so far
  • h
  • a
    cupcake dataset wip #2
  • a
    Huh. How much audio do you have so far?
  • a
    incase if someone wonders what software i used, it's called rx7, i used audacity but now i wanna try on rx7
  • a
    about lets say...
  • a
    165 so far
  • a
    holy crap
  • a
    in i think 3 episodes
  • a
    now doing 4th
  • a
    Then your model would turn out really good then
  • a
    and also that show was in netflix that actually has 5.1 audio
  • a
    which is great
  • a
    it has a bit noise of music and sound effects but i use dialogue isolation to do that
  • a
    I'm now gonna collect audio for Palutena, so then all 3 of the Kid Icarus characters can be made
  • a
    o, nice good luck getting it
  • a
    and training it
  • a
    I can't really train it, but I can make a dataset for it
  • a
    oh wait you made it already? oh, well hope someone could train it
  • h
    would 27 minutes worth of ripped voice clips from media featuring Snow White suffice?
  • k
    Yes most likely
  • u
    I'm always getting the K80 GPU with the Colab notebook, how long does it take to get a GPU that isn't a K80?
  • h
    How should I lay it out properly for training tho?
  • h
    I'm a little intimidated with the transcription part
1...118119120...1068Latest