You can use this to make images that split subjects down the middle with different designs on the left and right.

This was made to work with a broad range of subjects. If it doesn't work with your prompt, turn the weight up. It works well with people at lower weights, but other subjects may need much higher weights (up to 1.5). If you're using a long prompt and it doesn't work, try shortening the prompt down to the essentials.

Keep in mind that SDXL has limitations when prompting for multiple subjects. You can ask for certain features, but you can't guarantee you'll get both parts, what stays together, or what ends up on the left or right.

This LoRA was trained on multiple styles, so it can do both photo realism and painterly styles. Without guidance, it tends to lean toward a digital painting style.