https://uberduck.ai/ logo
Powered by
# machine-learning
  • u
    1 minute of data is plenty
  • t
    I thought it would be great, that's why. Plus I thought it would be used for an actual Uberduck model but I was wrong.
  • t
    At least I was taught how to do it though.
  • e
    Wait you can’t use it for a uberduck model?
  • t
    Nope!
  • m
    tortoiseTTS isn't related to Uberduck in any way
  • m
    it was just a cool TTS colab we found
  • u
    and their training process is completely private
  • e
    But how do you put it in uberduck
  • u
    you dont
  • e
    Then how did other guy who used it put it in uberduck
  • u
    he didint
  • e
    But it’s in uberduck rn
  • m
    if it says its made with tortoiseTTS, then they just used audio clips of the tortoiseTTS generation and made a Tacotron model with it
  • e
    Why is it like that
  • m
    what you mean?
  • e
    Why you gotta get the tts you just made and put voice clips of it for a other model of it
  • m
    because tortoiseTTS just makes audio files, not a model
  • e
    Im confused
  • m
    what is there to be confused about, tortoiseTTS isn't open source so we can't use it
  • e
    Because this dude said he used it and it’s helpful for characters with few lines
  • e
    So I thought you made models with it
  • m
    no, they just took the audio files generated and made a Tacotron model that is actually suitable for Uberduck
  • e
    Whatever man
  • u
    test of Tattletail
  • u
    translation: I am tattletail i became an ai voice
  • u
    also a test of suction cup man
  • t
    "Unsupported bit depth: the wav file has 24-bit data." i am such a good girl i am such
  • t
    i fucked up. i fucked up and i have to convert 230 audio files individually
  • h
    false
1...915916917...1068Latest