Let's talk about the XL system and XL prompts.

First of all, there are two major differences between SD and XL; the first is the resolution. While SD works with 512 x 512 blocks, XL works with initial 1024 x 1024 blocks. The second is the prompt system itself.

Okay, BEFORE ANYTHING else, understand that I'm not saying that Yodayo is bad, not at all!!! I love Yodayo, just as I love Yoda-chan.

However, XL works with two prompt systems. The primary and most important is LINGUISTIC POSITIVE and the second is SUPPORTING TERMS. Here on Yodayo, however, you can only work with one, unfortunately. So it's not possible to get the most out of XL (yet).

That said, the best way to create prompts for XL is through the context of the image you want to generate. This is where ChatGPT can come in handy. Also, as I've said before, XL is still in version 1, which means it still lacks a lot of features, and using LoRAs, for example from SD, doesn't work, as you can see in the image opposite. My intention was to make an image of Yoda-chan and well... 😅😅😅

Prompt: 1girl, solo, yoda-chan, bangs, long hair, paintbrush, multicolored hair, <lora:Yoda-Chan-10:0.8> Model: HoloDayo XL

It looks cute, but it's not what I wanted... (nor what I expected)

Anyway, if you want to reproduce an image on the XL but don't know where to start, take an image you like and use it here: https://huggingface.co/spaces/pharmapsychotic/CLIP-Interrogator It's not perfect, but it can give you an idea of what to do... That's it, all you have to do is create and enjoy... 🥰🥰🥰

Btw, I recommend using 70 steps, DPM++ 2m Karras (while Yodayo doesn't have the 3M Exponential)

Due to the 2k character limits per post, to see all 10 parts of the guide about prompts, go to my profile: https://yodayo.com/posts/01a5e690-eabf-45cc-8309-4d0f931b7178/

My Instagram: https://instagram.com/chaos.experience.ai