Draw the composition of your AI image

Hello Creative Crusaders

Stable Diffusion keeps changing the image creation landscape and animation feels withing reach!

So far I’ve show you how powerful ControlNET is for controlling composition and I want to introduce you to another extension for the same purpose, where you can draw your composition and assign prompt to each individual element.

It requires two extensions: composable lora and latent couple.

To install them, you can follow this guide. It also shows this technique in video format.

Once you have your Extensions in Stable Diffusion set up, you first need to enable both Composable Lora and Latent Couple.

The Latent Couple area is where we will draw our image. We can upload it or draw it directly in Stable Diffusion.

First we create a blank canvas and draw the composition of our image. Use different color for different elements. Press I’ve finished my sketch once you’re done.

We can write a general prompt for our scene as well as prompts for each individual element. We can adjust the weights for each of the elements. Hit Prompt Info Update when you’re done and it will automatically copy it to your prompt window.

Adjust your general settings and hit generate.

As usual, experiment with the settings, weights, images and prompts until you find something that works for your own use case.

I’ll be doing a live walkthrough of this and several other Stable Diffusion techniques this Thursday at 2pm GMT+7.

If you want to dive deeper into it, join me on the AICC discord server.

You can also stop by my website or social media and as always,

Keep creating and don’t forget to have fun. ☀