Think 8GB VRAM Can't Handle Controllable AI Generation? Naaaaaah! Introducing ControlNeXT SVD

Поделиться
HTML-код

Комментарии • 43

  • @nitsgamer7740
    @nitsgamer7740 Месяц назад +3

    Appreciate all these easy to follow tutorials.

  • @CalsProductions
    @CalsProductions Месяц назад +6

    It will be awesome if you can do a tutorial on your intro avatars, the lip sync is and head movement are great!

    • @StableAIHub
      @StableAIHub  Месяц назад +5

      I already did.
      ruclips.net/video/WtHdvSSQlWo/видео.html

    • @CalsProductions
      @CalsProductions Месяц назад +5

      ​@@StableAIHub Wow super cool, thanks!!!!!

    • @StableAIHub
      @StableAIHub  Месяц назад +4

      Np. The tutorial is a bit old. In case you face problems during installation, plz let me know.
      Let me know once you install it successfully, I will let you know the settings with which it produce good output. I did a lot of trial and error.

    • @CalsProductions
      @CalsProductions Месяц назад +5

      @@StableAIHub I have installed it and it works great, but not like your outputs.

    • @StableAIHub
      @StableAIHub  Месяц назад +5

      @@CalsProductions What is the actual issue. Is it that the output was bad and deformed or is it the quality. For quality improvement I already mentioned in demo page (after the intro) you will have to do Interpolation + upscale + gfpgan. I already posted video on interpolation yesterday. and other aspects in earlier videos (Diffsynth).

  • @americo9999
    @americo9999 Месяц назад +3

    I hate not having a good computer, what type of computer do you need? mine doesn't have a beefy GPU , is 10 years old ;'( and lastly, can this run in any Linux distro?

    • @StableAIHub
      @StableAIHub  Месяц назад +1

      What is the budget? NVIDIA GPU with 16 GB VRAM is good starting. You can also try on Google Collab

  • @baseerfarooqui168
    @baseerfarooqui168 Месяц назад +4

    is echomimic software can generates lipssync on video? thanks

    • @StableAIHub
      @StableAIHub  Месяц назад +4

      It does but I could never get it to work on Windows.

  • @trishul1979
    @trishul1979 Месяц назад +5

    Please share what did you use for increasing FPS of video. It worked fine on 12 GB VRAM and generated output but it's not that smooth.

    • @miycko
      @miycko Месяц назад +3

      ComfyUI has a variety of options for frame interpolation, they work great but my go to is actually capcut, they hve a slomo feature on their online video editor which you can use to precisely smooth out your video with customizable curves. its free and works wonders

    • @StableAIHub
      @StableAIHub  Месяц назад +4

      Sure will post quick tutorial

    • @StableAIHub
      @StableAIHub  Месяц назад +4

      @miycko Thanks for the information. I think he is from India (commented in one of the earlier videos) and Capcut is banned here (in India).

    • @miycko
      @miycko Месяц назад +5

      My pleasure, and thank you for your installation guides. It would be impossible to get through any of these without your hard work

  • @sandeepm809
    @sandeepm809 Месяц назад +3

    is there a ComfyUI workflow ??

    • @StableAIHub
      @StableAIHub  Месяц назад +2

      Sorry I don't use ComfyUI.

    • @crystian7016
      @crystian7016 26 дней назад +2

      @@StableAIHub You should. There's one version which cointains a built-in extension manager, so you don't need to install them step by step.

    • @StableAIHub
      @StableAIHub  26 дней назад +1

      @@crystian7016 I have used in past but since it is complex I gave up.

    • @crystian7016
      @crystian7016 25 дней назад +2

      @@StableAIHub When you don't have the apropriate workflow it is confusing, indeed, but when you have it prepared you just need to do one click to generate. And you can have many of them. There's some function where you can drag and drop an image and the software generates the workflow which created it. I couldn't use this function, but i will soon. The best part of this software is that you can directly associate outputs with another image processing tools. Very professional.

    • @crystian7016
      @crystian7016 25 дней назад +2

      @@StableAIHub I've started with the standard workflow. I could generate images with stable difusion and fooocus with no problems, but when i tried with Flux i couln't. I changed the workflow and i've setup some other things, but the problem was just with the "cfg" config, it appears that flux doesn't work very well with it at higher values, but between 1-2 it was fine. The standard was 8. In the cfg description it says that it's an option which at higher values it increases the context recognition power, but at the same time it lowers the image quality.

  • @marcusk7855
    @marcusk7855 Месяц назад +5

    Can't it do NSFW?

  • @CringeGPT-gnt
    @CringeGPT-gnt Месяц назад +4

    Be careful bro, you leaked your ip at 6:42

    • @StableAIHub
      @StableAIHub  Месяц назад +3

      Thanks for letting me know. It is remote server IP or am I missing something?

    • @CringeGPT-gnt
      @CringeGPT-gnt Месяц назад +4

      It is your public ip bro
      Be careful

    • @StableAIHub
      @StableAIHub  Месяц назад +3

      @@CringeGPT-gnt I just checked again, you talking about this IP, right
      whatismyipaddress.com/ip/52.89.167.149

  • @ApexArtistX
    @ApexArtistX Месяц назад +3

    The voice does not match the face

    • @StableAIHub
      @StableAIHub  Месяц назад +2

      Yeah it won't. Just to demonstrate the audio2LipSync capabilities.

  • @doomslayerforever2858
    @doomslayerforever2858 Месяц назад +4

    I HATE THIS

    • @StableAIHub
      @StableAIHub  Месяц назад +5

      What made you hate this?

    • @CoreyJohnson193
      @CoreyJohnson193 Месяц назад +4

      Aww, what's wrong, friend?

    • @StableAIHub
      @StableAIHub  Месяц назад +4

      @@CoreyJohnson193 Let me know once he tell you. ;)