Furry Distilled

This is an experimental furry-focused model distilled from Pony Diffusion v6 XL. It's essentially an attempt to shrink that model down and try to get similar results with better speed than the non-Turbo version and lower memory usage, much like Segmind's SSD-1B but less professional and more NSFW. This current release kind of works but doesn't really achieve that. If you're an end user who just wants good images I'd recommend sticking to Pony Diffusion XL as the better results easily make up for the speed difference.

Usage

You can use this in ComfyUI and possibly other programs. Generate images in 1024x1024 or other standard SDXL resolutions, using 25 or more steps and prompting with "score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, describe what you want, some tags" much like you would for Pony Diffusion XL. Using smaller image sizes or prompts not in this format will usually produce bad results. It's also not compatible with all the software that SDXL-based models work in.

Note that this was created by an amateur using buggy code, a fairly poor dataset, and not enough compute time. Don't expect the quality and level of prompt understanding or character knowledge of Pony Diffusion, or even the close-to-SDXL results of the original SSD-1B - much less resources went into this. It's likely that it could be improved substantially with a bit of effort.

Licensing

Since this is based on Pony Diffusion XL it's released under the same modified Fair AI Public License 1.0-SD which forbids inference of this model on websites or applications allowing any form of monetization (paid inference, faster tiers, etc.). This also applies to any derivative models or model merges.