Run ComfyUI Workflows In Open WebUI - Flux GGUF Workflow included - Local DALLE 3

Поділитися
Вставка
  • Опубліковано 22 гру 2024

КОМЕНТАРІ • 39

  • @Spestly
    @Spestly 2 місяці тому +3

    I got this error can some one help me please:
    Something went wrong :/ 'NoneType' object is not subscriptable

    • @joelandresnavarro9841
      @joelandresnavarro9841 2 місяці тому +2

      Were you able to fix it? (I'm using ComfyUI)

    • @lasdem
      @lasdem Місяць тому +2

      I get the same error, where you able to fix it?

    • @Klowhllectors
      @Klowhllectors Місяць тому

      same error, but i think it is because i use EasyStableDiffusion Docker image, and for it to work the Unet Folder should be in the defaut location (Maybe) mine is not in the defaut location.

  • @sudabadri7051
    @sudabadri7051 3 місяці тому

    Can this be done... not locally id love to share workflows with friends without gpus

    • @TheLocalLab
      @TheLocalLab  3 місяці тому

      I know Open-webUI supports some sharing features that could make this possible but I'm not sure how developed those features are as this comfyUI integration is relatively new. You can check there documentation webpage and see if it doable. - Link - docs.openwebui.com/features

  • @rd-cv4vm
    @rd-cv4vm 3 місяці тому +1

    hello, i have an issue with the ComfyUI Base URL, i put the right url (without dash at the end) and it refuses to connect, i added --listen to the comfy-gpu start. Would you know where the issue is from.Appart from that the rest is working thanks for the video

    • @TheLocalLab
      @TheLocalLab  3 місяці тому

      Can you paste the base URL here, the way you pasted it in open-webUI?

    • @rd-cv4vm
      @rd-cv4vm 3 місяці тому

      @@TheLocalLab 127.0.0.1:8080
      i changed the port from the run, there is nothing to do for me

    • @TheLocalLab
      @TheLocalLab  3 місяці тому

      I believe your using the wrong base URL. The base URL you pasted appears to be the URL for open-webUI. The comfyUI base URL normally ends like this 127.0.0.1:8188. Open- webUI runs on port 8080 and comfy 8188, at least in my setup.

    • @rd-cv4vm
      @rd-cv4vm 3 місяці тому

      @@TheLocalLab yes, i changed the port of comfy to match webui port , wishfully thinking it would help with the issue, the port change was done manually as it wouldn't work with the base URL you mentioned

    • @TheLocalLab
      @TheLocalLab  3 місяці тому

      It is not a good idea to have both projects running on the same port. This will certainly cause more issues then it solves. Try switching it back and check the comfy terminal to ensure the URL your using matches what's in the terminal before testing in open-webUI.

  • @atzha01
    @atzha01 3 місяці тому

    Hello, thank you for this, it is really helpful, got me wondering if this is something which can be adapted to be used for an image2image workflow or it is only suitable for text2image?

    • @TheLocalLab
      @TheLocalLab  3 місяці тому +1

      As the feature develops, possibly but I don't think there's a way to pass the images from open-webui to the comfy api just yet.

    • @atzha01
      @atzha01 3 місяці тому

      @@TheLocalLab Thank you - will keep an eye out for that - whenever it comes out

  • @TheLocalLab
    @TheLocalLab  3 місяці тому +1

    🔴Easy Open-WebUI + LM Studio Tutorial: Free & Local ChatGPT Alternative 👉 ua-cam.com/video/8u4SeFnY9Jc/v-deo.html
    👉 Want to reach out? Join my Discord by clicking here - discord.gg/5hmB4N4JFc

  • @abuhafiyyan9846
    @abuhafiyyan9846 3 місяці тому

    what GPU you are using?

  • @A_Little_O_Everything
    @A_Little_O_Everything 3 місяці тому

    This tutorial is great! I hope i word this where its understandable. I have a issue where I created a Comfy workflow where the image is upscaled and I don't have no preview image nodes and turned ksampler preview image off so I only get the final result. I used the saved api and when I prompt a image in open webui, it gives me 3 of the same image in which i would prefer 1 image. Do you have any suggestions?

    • @TheLocalLab
      @TheLocalLab  3 місяці тому

      You wouldn't have a copy of this workflow, so I can get a better understand of what you got going here.

  • @AnthonyPasturel
    @AnthonyPasturel 3 місяці тому

    Hello Local Lab, thanks for the video. I try to generate an image with llama 3.1 8b , but i have an error message "The size is not supported by this model." I ve tried to change the image's size with 256x256 in the images parameters (512x512 by defaut) but it didn't work. Did you already get the problem? Thanks by advance
    To be more precised it's with OPEN AI engine. Dall-e-3 model by default. And i've already added by API key.

    • @TheLocalLab
      @TheLocalLab  3 місяці тому

      Yeah Dalle-3 api doesn't support 256 x 256 resolution size. Try 10204 x 1024 instead.

  • @davisj789
    @davisj789 28 днів тому

    getting a port error even though its the right working port, any help?

    • @TheLocalLab
      @TheLocalLab  28 днів тому

      Port error for comfy? The default is 8188.

    • @davisj789
      @davisj789 28 днів тому

      @@TheLocalLab no in open web, it just says it doesnt exist even when copy stright from cmd

    • @TheLocalLab
      @TheLocalLab  28 днів тому

      @@davisj789 What's the url or port that your pasting?

  • @DeathMasterofhell15
    @DeathMasterofhell15 3 місяці тому +2

    About to make a movie