Quick How to! Use Stable Diffusion 3 in 3 minutes | Tutorial | ComfyUI | Colab | Stability AI API

I'll show you how to use stable diffusion or without comfy UI before we do anything else we need an API key so let's get that from the Stability AI website go ahead and log in or create an account if you don't already have one create an API key and copy it I'll show you how to use a free Google collab notebook open the collab notebook Linked In the description and make your own copy and open the secrets tab delete any keys here and click add a new secret you'll name this SD 3 API in all caps paste your key and allow notebook access now hit the run button for step 1 this will install all dependencies in the notebook once that's finished run step 2 and scroll to the bottom click on the link and this will take you to your very own stable Diffusion 3 generator update your prompts select your aspect ratio seed model style and then hit submit this will use the stability API to generate the image you'll get 25 credits to start with depending on the model you'll be able to generate three to six images the Turbo is cheaper but the full model definitely does better with text I created this notebook myself from another gitHub repo now let's do the same thing in Comfy UI go ahead and launch your Comfy UI instance go to the manager and install custom nodes you can search for stable Division 3 or Zho to see if it's available in the manager at this time it's not available so we'll install it manually using terminal open your terminal and change the directory to your custom nodes folder then clone the gitHub repository after it's finished open the folder and the configuration file and paste your stability AI API key between the quotation marks you'll now need to restart your comfy UI and refresh your browser search for the stable diffusion 3 node paste in your prompts and select your settings you can choose your aspect ratio the model you can fix your seed and set your denoise strength for the image and if you want to do text to image or image to image here I'm feeding in an image by loading it and you'll see that it has changed the aspect ratio let's do one more run let's create a new sign from text to image and see if it can handle multiple text outputs the full model definitely produces better results but it's far from being perfect and you'll most likely need a few tweaks and credits to get the image you're looking for and that's pretty much it let us know what you think of the new models in the comments below if you're out of the loop with what's going on at stability AI click here 3 to generate images with after lots of trial and experimentation

Channel: Abe aTech

Views: 1,684

Keywords: stability ai, stable diffusion 3, midjourney, tutorial, how to, guide, comfyui, comfyui tutorial, colab, colab tutorial, SD3, sdxl, stable diffusion, a1111, automatic1111, sd1.5, ai art, ai art generator, ai art community, image generation, dall-e 3, emad mostaque

Id: yRYAqubZzNY

Length: 3min 7sec (187 seconds)

Published: Tue Apr 23 2024