InvokeAI - Canvas Tips & Tricks #1

Поділитися
Вставка
  • Опубліковано 8 лис 2024

КОМЕНТАРІ • 43

  • @emadalzahid3394
    @emadalzahid3394 Рік тому +2

    Thank you for your efforts in these details. I wish you continued success. Explained in a beautiful style

  • @Mehdi0montahw
    @Mehdi0montahw Рік тому +1

    like dream

  • @saas-space
    @saas-space Рік тому

    Awesome AI! Keep going!

  • @ViratxDoodle
    @ViratxDoodle Рік тому

    Where is the Image to Image tab in Unified Canvas of version 3.1 ?

  • @GrauerRauch
    @GrauerRauch Рік тому

    Pleace more Tutorials.

  • @robertmagpizza
    @robertmagpizza Рік тому +1

    Thank you very much! I just started playing around a few weeks ago and this video is a great help! Amazing Tool!

  • @user-zs4iv3vf5m
    @user-zs4iv3vf5m Рік тому

    if one day invoke ai supported windows AMD GPU, sure I would tried it.

  • @DanielSchweinert
    @DanielSchweinert Рік тому +2

    Wow this is amazing! Keep up the good work! Just begun to use it yesterday and Im blown away. Training your own LoRA / LoHA inside of Invoke AI would be also a great addition.

  • @zenko4187
    @zenko4187 Рік тому +2

    Really looking forward to 3.0 and nodes

  • @Hijab_Diffusion
    @Hijab_Diffusion Рік тому

    is there any working colab or runpod available?

  • @Feelix420
    @Feelix420 Рік тому

    just installed the 3.0 beta version and i can't seem to upload pictures on the canvas. it says image uploaded but there's no image on canvas.

  • @向泽晨
    @向泽晨 Рік тому

    After successful download, in Mask mode, the paintbrush cannot be used. What is the situation

  • @rizla007
    @rizla007 Рік тому

    I am missing a lot of the options, including face restoration in my menu

  • @borisrusev9474
    @borisrusev9474 Рік тому

    So there is no way to guarantee design consistency?

  • @DylanDValentine
    @DylanDValentine Рік тому

    Im really confused on how to use Lora in invokeAI and I cant find any resources on anyone explaining it.

  • @zephirius
    @zephirius Рік тому +3

    my major problem with invoke is that the png info are not compatible with A1111.. And not matter how much i take care of copy pasting the info, they always get lost in the img to img or upscaling etc.. process. The folder system where the outpaintings are recorded is strange too and i cannot choose where to write my files. It becomes really fast a mess when you go back and forth between the different functions. Everything else is really fine. Happy to see that you continue to work on it, will be happy too to pay for an app version !

  • @luizburattini
    @luizburattini Рік тому +2

    I used it for three weeks and ended up going back to SD webui. I didn't find support, there are few tutorial videos like this one and I missed Controlnet's functionalities. Invokeai is beautiful, it's fast, but I'll wait a little longer.

    • @zephirius
      @zephirius Рік тому

      same. but there some much people on A1111 how could they keep up ? a little sad that they choose to not be A1111compatible. so i went back toOpenOutPaint

  • @RickBoiter
    @RickBoiter Рік тому

    Where did the HipsterFantasy+Scifi diffusion model come from??

  • @pn7735
    @pn7735 Рік тому

    Could you give some details on the Model you are using? I did not find a model called "HipsterFantasy+SciFi".....?

  • @Xeraser2
    @Xeraser2 Рік тому +4

    A tutorial for outpainting/expanding a generated image would be nice.

  • @gasanmalikov
    @gasanmalikov Рік тому

    Please add a dropdown where you can select "Invoke forever"

    • @hangry3102
      @hangry3102 Рік тому

      What is "invoke forever" ?

    • @gasanmalikov
      @gasanmalikov Рік тому

      @@hangry3102 The continuous generation of images.

    • @hangry3102
      @hangry3102 Рік тому

      @@gasanmalikov Aah. I don't know how that'd fare considering InvokeAI's main goal is stability, and countless generations doesn't sound like the most stable of features. Either way I'm sure there's a possibility for such an extension in the future.

  • @miguelnoname5975
    @miguelnoname5975 Рік тому

    Hello InvokeAI peeps,
    Will X/Y/Z plots ever be implemented, like the one in automatic1111?
    I like using it to test different models.

  • @bruno10505
    @bruno10505 Рік тому

    I can't install it from itch, it shows "An error occurred during installation
    Cannot read property 'build' of undefined
    InvokeAl - The Stable Diffusion Toolkit"
    Does anyone know how to solve it?

  • @LaughterOnWater
    @LaughterOnWater Рік тому

    where can we download the hipsterfantasy+scifi model?

    • @Dougy64
      @Dougy64 Рік тому +1

      It is not yet released

    • @LaughterOnWater
      @LaughterOnWater Рік тому

      @@Dougy64 Said in the voice of Zeus from _Clash of the Titans:_
      "Release the model!"

  • @date2077
    @date2077 Рік тому

    What about photorealistic paintings?

  • @grimbrovonzombie8334
    @grimbrovonzombie8334 Рік тому

    How do you add more samplers than the installed defaults. You seem to have more options here in the video?

  • @mrzackcole
    @mrzackcole Рік тому +1

    Underneath the eerie moonlight, dense foliage shrouds an enigmatic gathering of tribespeople. Their faces are masked by shadows as they chant in a haunting, melodic rhythm. The camera slowly zooms in, navigating the labyrinth of trees and leaves, as the hypnotic chanting grows louder. Their bodies sway in unison, surrounding a pulsating, glowing object at the epicenter of their circle. The ethereal light casts an otherworldly glow on the scene. As the camera approaches, the tribespeople part with a sense of reverence, revealing the source of their devotion. It is a Stable Diffusion model, its radiance illuminating the words etched on its surface - "HipsterFantasy." ... The tension hangs heavy in the air, hinting at the dark and dramatic journey that lies ahead.

    • @invokeai
      @invokeai  Рік тому +1

      In their minds, a vision of the future suddenly forms.
      With a sudden burst of movement, the tribespeople break into a frenzied dance, their limbs flailing with the grace of a thousand drunken flamingos. Their chanting morphs into a bizarre medley of indie folk songs, their voices blending seamlessly with the haunting howls of wolves in the distance. The scene has become an electrifying hipster ritual, their fervor fueled by the magical allure of the Stable Diffusion model.
      A figure adorned in plaid flannel and sporting a robust beard steps forward, wielding a wooden staff topped with a vintage record player. As he raises his arms, the spinning vinyl emits a hypnotic melody, further entrancing the gathering. The tribe's elder, a wise man with a face obscured by a kaleidoscope of tattoos and piercings, joins the circle, cradling a sacred mason jar filled with organic kombucha.
      The ritual reaches its zenith as the tribespeople form an intricate human mandala, their limbs entwining and writhing like a tangle of artisanal vines. The Stable Diffusion model pulsates in response, its glow intensifying, casting a surreal light upon the scene. A sudden hush falls over the gathering, their breaths held in anticipation.
      With a sharp crack, the object's surface shatters, unleashing a torrent of exquisitely handcrafted items: knitted scarves, leather-bound journals, and bespoke ceramic mugs. The tribespeople gasp, awestruck by the divine HipsterFantasy bounty that now showers upon them like an avalanche of obscure references and cold brew coffee.
      As the last of the artisanal treasures settle upon the forest floor, the tribespeople resume their hypnotic chant, their voices now infused with a newfound sense of purpose and identity. A lingering word hovers in their minds, resonating on their inner anticipation like a sounding bowl crafted from recycled plastics - "Soon".

  • @tomschuelke7955
    @tomschuelke7955 Рік тому

    so.. i was thinking of installing this, but read, my old gforce 1080 wouldnt do it?
    But this with controlnet, for architectural renders would be great.. is this allready integreated? and something to train it for your own style.. aaaaaaand.. to run stable diffusion also on your own computer..
    My architctural company would love something like this, but are afraid to use stuff directly in the net without knowing where and how the data are used..

    • @invokeai
      @invokeai  Рік тому

      Hey Tom - Controlnet is coming out soon, and we'll be hosting Invoke (including for a few architects!) as a service. You can sign up for our waitlist at invoke.ai

    • @tomschuelke7955
      @tomschuelke7955 Рік тому

      @@invokeai Wow.. that was fast... love your current Programm.. as a usual architect the installation seemed to be welll.. steep learning curve needed.. but.. maybe next week i can point autogpt out to install it for me.. 😂

    • @HenriqueFrezza
      @HenriqueFrezza Рік тому +1

      @@tomschuelke7955 I'm running InvokeAI with a 1080 on my production computer with no issues so far. Everthing works great, txt2img img2img, canvas, loras etc.

    • @tomschuelke7955
      @tomschuelke7955 Рік тому

      @@HenriqueFrezza thanks for. Your hint on this.. Ill try this when 3.x is available

  • @imlean
    @imlean Рік тому

    can you share the model used so we can produce similar results? thanks

  • @RamonGuthrie
    @RamonGuthrie Рік тому

    Can I set a LoRa folder path outside of InvokeAI to link to?

  • @Lansolot
    @Lansolot Рік тому

    ControlNet destroyed InvokeAI for me, really wish implementation for extensions was considered. It's a far superior interface but it's missing out on so many extensions that it makes it just a good interface.