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

    JacobLenstar

    05/16/2021, 2:25 AM
    as I’ve learned
  • c

    ccateni

    05/16/2021, 2:25 AM
    interesting, I'll take note.
  • c

    ccateni

    05/16/2021, 2:25 AM
    thanks for the help.
  • j

    JacobLenstar

    05/16/2021, 2:25 AM
    👍
  • l

    Lu9

    05/16/2021, 5:58 AM
    i am trying to train a model on 460 samples but when generating mel spectograms it only does 18/18 pretty much, and i've changed the batch size so idk whats happening
  • l

    Lu9

    05/16/2021, 5:58 AM
    after that it will look for npy files that dont exist
  • r

    redringofb

    05/16/2021, 6:04 AM
    stop posting about bfb im tired of seeing it
  • l

    Lu9

    05/16/2021, 6:06 AM
    could it be because some of the voice clips are shorter than 2 seconds or something else entirely
  • r

    RAYTRAC3R/Dhama the Llama (Sami)

    05/16/2021, 6:14 AM
    ???
  • l

    Lu9

    05/16/2021, 7:19 AM
    well, i was sure all of my wavs were 16 bit 22050 Hz but i decided to re-encode them either way, and now its working...
  • l

    Lu9

    05/16/2021, 8:22 AM
    any of the current collabs used by the community have support for hifi gan? just curious
  • u

    user

    05/16/2021, 8:29 AM
    Why? I Like BFB
  • p

    pi

    05/16/2021, 11:00 AM
    can I see your dataset?
  • p

    pi

    05/16/2021, 11:00 AM
    oh ok
  • z

    Zachary(retired f/ voice clonin)

    05/16/2021, 11:07 AM
    Sorry but i'm an BFB fan.
  • k

    kitsuner

    05/16/2021, 4:24 PM
    I say try to make a model anyway and just have fun with a limited vocabulary. That sounds like a fun challenge
  • c

    ccateni

    05/16/2021, 4:26 PM
    I'll try, but doing some research should help make a better model. especially with the time it takes to make one.
  • j

    Justin

    05/16/2021, 4:32 PM
    Again..audio credits: @User
  • u

    user

    05/16/2021, 5:13 PM
    His mouth is wide
  • b

    brilliam.de

    05/16/2021, 7:35 PM
    How did you get that working?
  • p

    pi

    05/16/2021, 7:40 PM
    https://www.nvidia.com/de-de/omniverse/apps/audio2face/
  • p

    pi

    05/16/2021, 7:40 PM
    installing and then load an audio file
  • p

    pi

    05/16/2021, 7:41 PM
    only works with rtx gpus though
  • b

    brilliam.de

    05/16/2021, 8:06 PM
    Interesting
  • t

    Toasty

    05/16/2021, 8:42 PM
    hmmm
  • t

    Toasty

    05/16/2021, 8:42 PM
    im tempted in attempt making a model on a chromebook
  • m

    mega b

    05/16/2021, 8:42 PM
    I'm making a youtube -> dataset maker
  • m

    mega b

    05/16/2021, 8:42 PM
    you input a youtube url and outcomes a dataset fast
  • b

    brilliam.de

    05/16/2021, 8:42 PM
    oh thats super cool
  • m

    mega b

    05/16/2021, 8:43 PM
    this should allow chromebook users to make models fast
1...888990...1068Latest