This is a quick and rough checkpoint merge I made based on AutismMix_confetti and some style loras. It's just based on personal preference and an experiment for my first merge. I'm not 100% satisfied with it but shared it just in case anyone wants to try it out. I will maybe update it sometime in the future,but for now check it out if you're curious and want to fill up your hard drive with extra AI junk.

Instructions:

1. Use standard pony logic and style tags (score_9, score_8_up, score_7_up, score_6_up,etc)

2. Use any sampler you want,I generally use Euler a Automatic/Euler a Align Your Steps/DPM ++2M Karras at 20-40 steps

3. CFG at 7-9

4. I recommend this Adetailer model for faces

5. I recommend using Reforge instead of A1111

Images were made at standard PDXL resolutions (1024x1024,832x1216,1216x832) with Adetailer,and then upscaled 2x.

And most importantly...

IF YOU USE STYLE LORAS ON A CHECKPOINT THAT IS ALREADY A MERGE OF STYLE LORAS,IT DEFEATS THE PURPOSE.

ok thanks bye