AI images
- Thread starter pwe
- Start date
-
Guest, did you like the last podcast notification?? Well here's the second one, and for the third podcast! Yep. This Thursday 8-8-24 at 7pm Central
Go here to check out his podcast thread for info on topics and links!Go into the FoH discord and find the Happy Time Fun Hour channel and give it a listen!
elcaro1101
Ahn'Qiraj Raider
- 2,083
- 4,958
Gutterflesh
Parody Account
- 8,127
- 42,361
I happened to stumble across Image Creator in Paint.
Started poking around with goofy stuff like goats playing poker, scenes from Blade Runner, cover art from Grand Theft Auto, a space cowboy, gamer eating Doritos etc. All of which generated without issue. I never touched on anything political, LGBTBBQ+ or sexual in any way.
What I was never able to generate was an image that included the word "knight". I kept removing words like "battle field", "dark", "grave" with the final prompt being watered down to "A knight kneeling in a gloomy field" which got me blocked.
Started poking around with goofy stuff like goats playing poker, scenes from Blade Runner, cover art from Grand Theft Auto, a space cowboy, gamer eating Doritos etc. All of which generated without issue. I never touched on anything political, LGBTBBQ+ or sexual in any way.
What I was never able to generate was an image that included the word "knight". I kept removing words like "battle field", "dark", "grave" with the final prompt being watered down to "A knight kneeling in a gloomy field" which got me blocked.
- 2
- 1
The open source Flux release you can run locally and is pretty good with text, so you can get a good sense of how the image and the company name might interact. Has pretty good variety, and you can tell it things like "use simple vector shapes" to control the design parameters.
When I played around with it, I used ComfyUI for the interface. If you have Nvidia GPU, there's a standalone windows release that's as simple as dragging a folder onto your desktop.
It can take a bunch of different models. Instructions for getting Flux running are here. You drop the model files in the appropriate folders, go into ComfyUI, and "load" the sample image on the page which also has a prebuilt workflow that you can start prompting against. If you've got a slower GPU, you can queue up a bunch of prompts then come back later to check the results.
Below are a few examples for the same prompt, so it has pretty good variety in style/interpretation (apparently it isn't very sneaky when it comes to pineapples, though):
Company logo. Company name is "Bobcat knife sharpening". Use simple vector shapes. incorporate knife sharpening into the design. Add a small, subtle pineapple into the logo subtly somewhere.
Removing the simple shapes bit to show more detail
- 2
- 1
AVTB
Vyemm Raider
- 180
- 2,115
Bandwagon
Kolohe
- 23,785
- 63,502
AVTB
Vyemm Raider
- 180
- 2,115
So that falls into the same pattern I was getting on the Facebook AI. They were getting progressively more cartoonish and picture or meme ready vs a logo he could put on a patch/stickers/freebies. I really haven’t played with any of the other platforms and don’t want to fire up a paid subscription as I won’t use it for anything else at the moment, but I’m not doing it by hand for sure
.I really appreciate the help and responses. I passed them to him so he can explore them and tweak the request on his own.
- 1
