Type | |
Stats | 54 5 |
Reviews | (6) |
Published | Dec 13, 2024 |
Base Model | |
Hash | AutoV2 86CEB59D37 |
First attempt at a Flux Lora! Why use Flux? Because it's the only model I know that seems to do text fairly accurately, and that's pretty important for this Lora.
This LORA has been trained on the older Enhanced Edition "cover" art of Sentinels of the Multiverse character. I find the goofy, amateur look of the art there more charming than the more polished-looking and modern Definitive Edition. The LORA can do proper title text correctly most of the time, usually only messing up by one or two letters otherwise.
Use natural language, and take a look at the sample images to see what the prompting format is like. Images should be Portrait Resolution (832x1216) - here's a breakdown of the format (do it all in one "sentence" rather than separate paragraphs)
A comic book cover featuring a character named "X" On the cover, the character has (write your char description here)
The background depicts (write your background description here)
The title "X" is prominently displayed at the top in bold, stylized letters with a Y theme. This comic is published by "Sentinel Comics" and is issue number 30, as indicated by the publisher's logo and the issue number on the cover.
You can put any number in the place of 30.
If you wish to add a subtitle to the title - for example, "The Amazing" "Spooder-Man", do it like this:
The title "Spooder-Man" is prominently displayed at the top in bold, stylized letters, with the subtitle "The Amazing" above it.
However, I noticed that whether or not it understands "above" or "below" is completely up in the air, so you'll probably have to roll the gacha a bunch of times.
I don't know whether Flux knows popular anime characters etc - if you want to try using this with characters loras (if they exist), please do so. Otherwise just generate as much as you can and please upload them so I can see the results of your creativity!