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

    CDS

    03/03/2023, 3:51 PM
    last model i trained was on a k80 and it worked just fine with 8
  • j

    Jellirby

    03/03/2023, 4:44 PM
    It's good to know that your components aren't a bottleneck.
  • t

    Tadejro

    03/03/2023, 7:19 PM
    Try my tool on your dataset: https://discord.com/channels/768215836665446480/994486394049282058/1079564245211033671
  • j

    Jellirby

    03/03/2023, 9:46 PM
    How'd this work with google collab. I'm curious since it should all be in google's end
  • u

    (".Lukus_bt.")

    03/03/2023, 10:04 PM
    help please
  • j

    Jellirby

    03/03/2023, 11:00 PM
    the dataset I think is fine. it's just more an issue with the transcript I think
  • j

    Jellirby

    03/03/2023, 11:16 PM
    this warning seems to come up a lot. maybe I should've turned down the volume on the dataset...
  • g

    Gosmokeless28

    03/04/2023, 12:27 AM
    Please use this notebook instead of that one: https://colab.research.google.com/drive/1p0Vh_GFiYQ84ECCrjU2PDkmhKcN-GEfR
  • u

    (".Lukus_bt.")

    03/04/2023, 1:45 AM
    ...
  • j

    Jellirby

    03/04/2023, 1:46 AM
    Patience, young one. Help will come eventually
  • g

    gabrielck11

    03/04/2023, 3:36 AM
    I do use that but the voices always sound jibberish
  • t

    Twilight Sparkle

    03/04/2023, 4:46 AM
    /content/project <>:177: SyntaxWarning: 'int' object is not callable; perhaps you missed a comma? <>:177: SyntaxWarning: 'int' object is not callable; perhaps you missed a comma? :177: SyntaxWarning: 'int' object is not callable; perhaps you missed a comma? speaker_count = 1([int(i.split("|")[2]) for i in lines]) + 1 --------------------------------------------------------------------------- IndexError Traceback (most recent call last) in 175 with open(target_filelist) as f: 176 lines = f.read().splitlines() --> 177 speaker_count = 1([int(i.split("|")[2]) for i in lines]) + 1 178 wav_count = len(lines) 179 in (.0) 175 with open(target_filelist) as f: 176 lines = f.read().splitlines() --> 177 speaker_count = 1([int(i.split("|")[2]) for i in lines]) + 1 178 wav_count = len(lines) 179 IndexError: list index out of range
  • m

    Mann Maker Studio

    03/04/2023, 8:07 AM
    can someone help with the synthesis notebook ?
  • g

    Gosmokeless28

    03/04/2023, 8:10 AM
    Is that the legacy TT2 synthesis notebook?
  • m

    Mann Maker Studio

    03/04/2023, 8:14 AM
    yes
  • g

    Gosmokeless28

    03/04/2023, 8:15 AM
    Summoning @Justin
  • j

    Justin

    03/04/2023, 9:00 AM
    not my nb
  • g

    Gosmokeless28

    03/04/2023, 9:22 AM
    Oh
  • g

    Gosmokeless28

    03/04/2023, 9:22 AM
    @Mann Maker Studio Are you using the wrong notebook?
  • m

    Mann Maker Studio

    03/04/2023, 9:24 AM
    no, it's from here https://colab.research.google.com/drive/1NVA3ndxhYWsKn-zwh3NnzMMgoVdJ5xUx
  • g

    Gosmokeless28

    03/04/2023, 9:41 AM
    Then that's the wrong notebook.
  • m

    Mann Maker Studio

    03/04/2023, 9:45 AM
    where is the actual one ?
  • g

    Gosmokeless28

    03/04/2023, 9:45 AM
    in #841437191073955920
  • m

    Mann Maker Studio

    03/04/2023, 10:04 AM
    can someone convert the Link2Model to accepts the .wav file ? https://colab.research.google.com/drive/1mzqy4Pb_yqIOkL4RfvuYsni4QXXdO_EN?usp=sharing#scrollTo=ODI651CWUdFs
  • g

    Gosmokeless28

    03/04/2023, 10:35 AM
    It only accepts YouTube videos because that's what it's supposed to do.
  • m

    Mann Maker Studio

    03/04/2023, 10:36 AM
    oh
  • s

    Sonic2022_mario

    03/04/2023, 2:17 PM
    I Made A Toopy Model, But Gosmokeless Is Offline, Can Someone Approve This Bot Command? toopy
  • s

    Sonic2022_mario

    03/04/2023, 4:54 PM
    It Also Needs To Work With TensorFlow2
  • g

    gabrielck11

    03/05/2023, 11:00 AM
    Im trying to make a Pipeline model and all of a sudden it stops training (it doesnt give me error)
  • t

    TheWu-Cepticon

    03/06/2023, 11:43 PM
    I'm new to using the Uberduck Pipeline training notebook. How do you know when you have the best possibly trained model?
1...127128129...158Latest