https://uberduck.ai/ logo
Join Discord
Powered by
# tacotron-2-support
  • g

    Gosmokeless28

    05/12/2023, 7:29 AM
    Then that's allowed to be on uberduck.ai
  • i

    Irsen Tanaka

    05/12/2023, 7:30 AM
    The same voice is for Chip-chan Kick! The Audio novel.
  • i

    Irsen Tanaka

    05/12/2023, 7:30 AM
    ydlmz just made his voice more closer to SatAM Robotnik.
  • y

    YTR76

    05/12/2023, 9:29 PM
    How much lower in percentage should the batch size be for wavs?
  • g

    Gosmokeless28

    05/13/2023, 3:50 AM
    What does that question even mean?
  • y

    YTR76

    05/13/2023, 3:38 PM
    What’s the recommended batch size for any amount of wavs in the dataset?
  • g

    Gosmokeless28

    05/13/2023, 6:02 PM
    Well, batch size cannot be higher than the number of wavs in a dataset, so there is no "one size fits all" batch size in all cases.
  • y

    YTR76

    05/13/2023, 6:02 PM
    oh ok
  • a

    Amizade | Pony's voice creator

    05/13/2023, 6:11 PM
    The egghead is not training so well. What did I do?

    https://cdn.discordapp.com/attachments/994486394049282058/1107007119284252773/image.png▾

  • a

    Amizade | Pony's voice creator

    05/13/2023, 6:14 PM

    https://cdn.discordapp.com/attachments/994486394049282058/1107007963501170688/image.png▾

  • a

    Amizade | Pony's voice creator

    05/13/2023, 6:14 PM
    Almost training well, but the egghead is flickering
  • a

    Amizade | Pony's voice creator

    05/13/2023, 6:25 PM
    @Cris140 I managed to make the voice here, but it's too echoed. Why is it like this? https://cdn.discordapp.com/attachments/994486394049282058/1107010785164677173/download_5.wav
  • c

    Cris140

    05/13/2023, 6:29 PM
    Problems on the audios
  • c

    Cris140

    05/13/2023, 6:29 PM
    There must be clean voice clips
  • a

    Amizade | Pony's voice creator

    05/13/2023, 6:30 PM
    The validation loss had lowered to 0.170 and returned to 0.200
  • a

    Amizade | Pony's voice creator

    05/13/2023, 6:30 PM
    I managed to isolate the voice, but how do I clean the voice?
  • a

    Amizade | Pony's voice creator

    05/13/2023, 6:30 PM
    On audacity?
  • s

    Sonic2022_mario

    05/13/2023, 6:33 PM
    Input?
  • a

    Amizade | Pony's voice creator

    05/13/2023, 6:35 PM
    Cris said that the problem is in the audios, but I isolated the voice and I don't know how I clean the voice
  • c

    Cris140

    05/13/2023, 6:40 PM
    Show samples of the dataset
  • a

    Amizade | Pony's voice creator

    05/13/2023, 6:41 PM
    okay, I'll look for a colab where I can clean the voice
  • a

    Amizade | Pony's voice creator

    05/13/2023, 6:42 PM
    but I made 26 wav audios here and I don't know which of these audios have noise that disturbs the voice
  • c

    Cris140

    05/13/2023, 6:45 PM
    send one sample
  • a

    Amizade | Pony's voice creator

    05/13/2023, 6:46 PM
    I have a better idea, I'm going to compress the audios I have in ZIP and I'm going to send them here
  • a

    Amizade | Pony's voice creator

    05/13/2023, 6:47 PM
    @Cris140 https://cdn.discordapp.com/attachments/994486394049282058/1107016198790578287/Audios_wav.zip
  • s

    Sonic2022_mario

    05/13/2023, 6:50 PM
    I Mean What Text You Putted
  • c

    Cris140

    05/13/2023, 6:51 PM
    Yeah it's probably a problem with the transcription and the audios
  • a

    Amizade | Pony's voice creator

    05/13/2023, 6:52 PM
    And what's the problem with the transcription?
  • a

    Amizade | Pony's voice creator

    05/13/2023, 6:53 PM
    So if the problem is the transcription, I'll have to look for a very useful collab to fix the transcription
  • c

    Cris140

    05/13/2023, 6:58 PM
    No, check manually
1...154155156157158Latest