https://uberduck.ai/ logo
Join Discord
Powered by
# tacotron-2-support
  • i
    The pronunciation will be fine, if the samples quantity will be enough. I am gonna add more samples.
  • a
    Okay. Good luck
  • y
    training the model in pipeline with aitch makes it more flexible and quick to train
  • a
    Next weekend, I'll try to use a pipeline.
  • s
    Bluey's Official Actor Is Amanda Moor And Bingo's Official Actor Is Chloe Pearson (According To Gabherelol)
  • i
    OK, I started over and tried to add some samples. They are as clear as the spring water, but the graph is still not good. Can You please look and try to help? https://drive.google.com/file/d/1vhylYIheaSsze1tw9e7oKO6oyYkzCfGS/view?usp=share_link
  • i
    No, it is actually an impression.
  • i
    Here is the result. For legacy. The pipeline output is even worse.

    https://cdn.discordapp.com/attachments/994486394049282058/1107918598141399060/mnNgnQAAAAASUVORK5CYII.png▾

  • i
    And what I have got at Pipeline:

    https://cdn.discordapp.com/attachments/994486394049282058/1107920093217824778/D0P9iZpmqI9kAAAAAElFTkSuQmCC.png▾

  • i
  • a
    This weekend I'll check it. Because I don't use my PC on weekdays
  • i
    I'll send You all iterations' models. But even if the eggheads graphs are good, model may fail. I have sent You samples and transcriptions list, so You can help me to fix the model.
  • a
    Okay, but I'm learning yet
  • s
    I Will Be Doing Profily's Voice From BFB, And These Are The Samples I Will Be Taking It From:

    https://www.youtube.com/watch?v=VIizPCqZcZI&pp=ygULYmZiIHByb2ZpbHk%3D▾

  • s
    Darn It, Can't Download The Vocals On Media.io

    https://cdn.discordapp.com/attachments/994486394049282058/1107984638061846529/Screenshot_2023-05-16_065453.jpg▾

  • p
    use a different website for downloading audio
  • g
    vocalremover.org
  • s
    The Audio Is 10 Minutes And 14 Seconds, It Won't Fit I Think
  • t
    Download Bandicam and use the audio only option, then split the vocals into short voice clips. MVSEP also helps remove background music
  • i
    If You know how to use Audacity, You can just split tracks by "Export splitted files" or "Export highlighted...".
  • h
    There is something wrong with the Pipeline notebook while i trained the Donut model. i mean, i did everything right!

    https://cdn.discordapp.com/attachments/994486394049282058/1108136841846194236/Screenshot_2023-05-16_165816.png▾

  • c
    No
  • c
    One or more audios are in stereo
  • s
    I Have Just Finished The List For Yellow Face's BFB Voice, The One That Is On The Site Sounds Like BFDIA

    https://cdn.discordapp.com/attachments/994486394049282058/1108357489910628362/Screenshot_2023-05-17_073610.jpg▾

  • s
  • a
    is tensorboard broken?

    https://cdn.discordapp.com/attachments/994486394049282058/1108505096427417722/image.png▾

  • a
    DUDE
  • a
    FREE SCAM HERE
  • a
    @HELP MODS
  • a
    he's gone yay
1...154155156157158Latest