The first version of the Flux model that I attempted to create. Please leave your feedback if you have any suggestions.
Due to hardware limitations, I can only provide the FP8 version.
Type fp8_e4m3fn
Description
FAQ
Comments (15)
any chance of Schnell?
Is there any positive& negative prompt suggestions, please?
just use natural language. positive& negative not necessary.
Flux uses natural language. While you can use SD style tags, negative prompts are typically ignored by Flux, or they can negatively impact your output. Just describe what you want, Flux will handle the rest.
Is NSFW available?
In ComfyUI, I always get the message: ERROR: Could not detect model type of
choose fp8_e4m3fn
@WAI0731 Where can I do that? Ich tried it with the file name, did not work.
@Miyuki_maho Image posted by WAI0731 (civitai.com)
@WAI0731 Thanks for help. It worked now. But the last problem is, what VAE could I use? The model loader dont have an VAE output.
@Miyuki_maho just use flux vae
I've been using your illustrious model along with my own Lora to create comics lately. The good news is that sometimes I can produce decent panel layouts, but it's still quite challenging. Hope you can incorporate more manga & doujinshi related materials into future flux training sessions.
If you want to use the current AI to draw comics, you might as well draw them yourself.
Comics are not about stitching together images. Composition, storyboards, narrative coherence, scene angles, and character presentation all need to be re-understood.
Current AI is a diffusion technology. Even flux can only understand part of natural language. To put it bluntly, it is actually stupid.
If a universal model for AI painting appears, it will be worth trying.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.