https://uberduck.ai/ logo
Join Discord
Powered by
# machine-learning
  • g

    GlobalIndustrial

    10/22/2022, 1:38 PM
    i tried quantizing the hifigan model and have created horrors previously unknown to man
  • h

    hecko

    10/22/2022, 1:59 PM
    you could sell that as a vst for $50
  • h

    hecko

    10/22/2022, 1:59 PM
    for extra horrors try setting speaker strength to like 10 in the synthesis notebook
  • i

    I love the hive

    10/23/2022, 10:01 AM
    he try to fall but dont know how to fall
  • j

    Justin

    10/23/2022, 1:31 PM

    https://youtu.be/-lz30by8-sU▾

  • e

    Epic toad

    10/23/2022, 11:51 PM
    Good
  • m

    moomoo

    10/25/2022, 1:57 AM
    are there any other singing model makers that give better quality than talknet ?
  • h

    hecko

    10/25/2022, 2:00 AM
    there's nnsvs but it requires very manual data work, beyond just transcription
  • u

    {K EY1} (Kei)

    10/25/2022, 2:03 AM
    If you need any help with nnsvs, ask me
  • m

    moomoo

    10/25/2022, 2:10 AM
    okay thank you so much ♥️♥️
  • u

    {K EY1} (Kei)

    10/25/2022, 2:13 AM
    This tutorial might help https://docs.google.com/document/d/1uMsepxbdUW65PfIWL1pt2OM6ZKa5ybTTJOpZ733Ht6s/edit
  • p

    postmates!!

    10/25/2022, 3:14 PM
    just put all my spamton clips into tortoise and this actually fits him alot
  • p

    postmates!!

    10/25/2022, 3:17 PM
    SPAMTON NO DONT SAY THAT
  • m

    moomoo

    10/25/2022, 3:37 PM
    thank you so much, you're an angel for real ♥️♥️
  • u

    {K EY1} (Kei)

    10/25/2022, 4:04 PM
    Np!
  • a

    Aidan, the owner of Goldrush!

    10/25/2022, 5:11 PM
    That's really what he sounds like!? That is so funny! Imagine a conversation between him and Mario., that is, Both of them in SM64, and just talking. lol
  • j

    Justin

    10/26/2022, 4:37 AM
    https://github.com/facebookresearch/encodec
  • j

    Justin

    10/27/2022, 6:40 AM
    https://github.com/chomeyama/SiFiGAN
  • g

    Gosmokeless28

    10/27/2022, 6:44 AM
    Exactly what is SiFi-GAN?
  • j

    Justin

    10/27/2022, 6:45 AM
    A fast and pitch controllable high-fidelity neural vocoder 😉
  • j

    Justin

    10/27/2022, 6:45 AM
    Demo: https://chomeyama.github.io/SiFiGAN-Demo/
  • j

    Justin

    10/27/2022, 12:38 PM
    https://github.com/MTG/Podcastmix
  • j

    Justin

    10/27/2022, 11:14 PM
    https://github.com/prophesier/diff-svc
  • i

    I love the hive

    10/29/2022, 7:47 AM
    I still try to make russian tacotron2 model but tacotron 2 make a english accent Text:Hello this is a russian model.
  • i

    I love the hive

    10/29/2022, 8:17 AM
    and same audio but in russian Text: Этот текст на русском
  • g

    Gordon4James5 Percy6 (Backup)

    10/29/2022, 4:29 PM
    Where Can I Upload Voices I Forgot
  • h

    hecko

    10/29/2022, 4:34 PM
    https://app.uberduck.ai/submit
  • i

    I love the hive

    10/31/2022, 9:43 AM
    Pls help I decided to train a HiFi-GAN but in generate mel spectrograms was given this error
  • h

    hecko

    10/31/2022, 10:03 AM
    you might need to change the symbol set
  • h

    hecko

    10/31/2022, 10:03 AM
    what language did you train the model in
1...9989991000...1068Latest