ComfyUI on Mac - step by step tutorial - AMD GPU or M1 M2 M3 Apple Silicon

Поделиться
HTML-код
  • Опубликовано: 11 ноя 2024
  • Apple Silison or AMD GPU - run stable diffusion ComfyUI on Macbook (with AMD GPU or M1 M2 M3). #comfyui #stablediffusion #macbooks #apple #applesilicon
    👉ⓢⓤⓑⓢⓒⓡⓘⓑⓔ
    👉 !! Try my online FLUX model free at agireact.com !!
    👉 updates: For Apple M processors, see • Apple Silicon run Fooo...
    The Git and Ananconda/Miniconda (for python virtual environment) are needed. Please make sure to install them first (for video tutorial, see • Macbook GPU (AMD or M1... ).
    ComfyUI homepage: github.com/com...
    Stable Diffusion 3 on ComfyUI: • Mac users: Stable diff...
    Installation commands:
    conda create -n fooocus python=3.10.12 -y
    conda activate fooocus
    pip3 install --pre torch torchvision torchaudio --extra-index-url download.pytor...
    git clone github.com/com...
    cd ComfyUI
    pip install -r requirements.txt
    #start UI
    python main.py
    Affiliate links: buy hardware on Amazon
    AMD GPU: amzn.to/3vCp6h1
    4600G amzn.to/45LhGFa
    5600G: amzn.to/3LgnFtC (same iGPU, better CPU)
    5700G: amzn.to/3Z9gUiM (better iGPU, and better CPU)
    ssd drive: amzn.to/3MVJdg2
    DDR4 drive: amzn.to/3sKNufi
    AM4 motherboard: amzn.to/3GfrPit
    PSU (power supply unit): amzn.to/3Gd87UA
    PC Case: amzn.to/3QPDNnF
    if you are interested in discrete GPU: amzn.to/3QT1wDp
    Please join the discord server at / discord where you developed free chatgpt bot and stable diffusion bot!
    If you would like to support me, here is my Kofi link: ko-fi.com/tech... and Patreon page: www.patreon.co...
    Thank you for watching!

Комментарии • 19

  • @AaronPaulRogers
    @AaronPaulRogers 6 часов назад

    Thanks brother!!!!! Great job

  • @sleepyblue8
    @sleepyblue8 2 месяца назад

    BRO thank you so much. I am a complete novice when it comes to anything dealing with terminals and coding. It all goes over my head. This tutorial was concise and very easy to follow. I appreciate you!

  • @Henyckma
    @Henyckma 2 месяца назад +1

    Thank you very much, i'm trying it out, new sub!

  • @saschas.4270
    @saschas.4270 2 месяца назад

    worked fine, but how do I restart it after a reboot?

  • @ywueeee
    @ywueeee 6 месяцев назад +2

    have you tried running animate diff, what's the usual processing times?

    • @tech-practice9805
      @tech-practice9805  5 месяцев назад +1

      for animatediff, I made several videos, ruclips.net/video/27M3xFZptmU/видео.html and ruclips.net/video/hTWU3JRB-N0/видео.html they take longer time as output is video (more frames)

  • @ropr433
    @ropr433 Месяц назад

    Thanks for the video!! I installed it, but it gives an error: "Trying to convert Float8_e4m3fn to the MPS backend but it does not have support for that dtype." - please tell me, maybe I need to update Pytorch? how do i update it? is there a command for this? I have a version now 2.6.0

    • @tech-practice9805
      @tech-practice9805  Месяц назад

      Hi! Please try: pip install torch==2.3.1 torchaudio==2.3.1 torchvision==0.18.1 you may also needs to update your OS system to latest if that doesn't work

  • @walterraponi
    @walterraponi 2 месяца назад

    i get a KSampler
    MPS BFloat16 is only supported on MacOS 14 or newer error, anyone resolved it?

    • @tech-practice9805
      @tech-practice9805  2 месяца назад

      what's your pytorch version? can try to upgrade it to latest

  • @InterestingLifeTravels
    @InterestingLifeTravels 3 месяца назад +1

    I’VE DONE THIS SO MANY TIMES AND I DONT HAVE IT??? PLEASE HELP. MISSING tensor 1 device=mps0 4:46

    • @saschas.4270
      @saschas.4270 2 месяца назад

      after mps device not found you have to click enter twice after that and then you can continue at 4:46. The ENTER is important, schau dir das Video genau an bei 04:34

  • @margaretthedy
    @margaretthedy 4 месяца назад +1

    when I try to generate the image I got blank image or damage image do you know why?

    • @tech-practice9805
      @tech-practice9805  4 месяца назад

      Hi! What GPU are you using?

    • @margaretthedy
      @margaretthedy 4 месяца назад

      @@tech-practice9805 thank you for your reply, my GPU AMD Radeon Pro 5700 8G

  • @ComfyUIstudio
    @ComfyUIstudio 2 месяца назад +1

    Nvidia 3090 makes this image in 3 seconds.

    • @yourenotallowed7494
      @yourenotallowed7494 2 месяца назад

      Yes. Performance will be always better as all these AI programs are optimized with NVIDIA CUDA cores, as well as 3D programs.

  • @TodayIFeelLike
    @TodayIFeelLike Месяц назад

    Dude, I've tried to watch several of your videos but you keep on assuming that people have seen your earlier videos if you're making a tutorial walk-through it with a step-by-step don't assume we've seen earlier videos because likely we have not and we're trying to learn everything in this video that you're showing. I was trying to watch your stuff, but I'm so frustrated by the fact that I would start and then at some point it was always watch earlier video. If you don't know what you're doing here that I just gave up and found another video so please for the future if you're making tutorials don't say referred to an earlier video just show us how to do it even at the videos a little bit longer show us how to do everything that we need to do for that particular function if we are showing us how to install com UI show us everything every single step from beginning to end.

    • @tech-practice9805
      @tech-practice9805  Месяц назад

      Thank you for the feedback! It needs python virtual environment install. You can either search for 'anaconda/miniconda install' or see my video ruclips.net/video/uOCL6h9fuVc/видео.html