This Model Increases/Decreases Detail

Detail tweakers allow for a more fine-tuned approach to adding and removing detail, instead of having to rely on prompting and janky weighted tokens to get the effect you need.

Inspiration

I generally really enjoy controlling the level of detail an image has (I prefer the negative direction) so I trained a model for pony diffusion initially, and it was brought to my attention that the SeaArt Furry XL model exists and in my experience thus far, this model is far superior to the Pony Diffusion checkpoint, it just requires a bit more ahem "finesse" to get your prompts right (something I will be experimenting in fixing in the near future)

Donations

Training models is expensive and I run training on my own private server, if you like what I do, consider supporting the development here!

https://ko-fi.com/yolup

Quirks

V1.5

• Effectiveness seems to be dependent on the artists used.

• Occasionally makes an image incoherent.

• seems to affect shading more than I would like

I plan on fixing the above issues, but I haven't slept in 18 hours and I kinda promised I would get something up for it in this time frame so here it is. I will be working on this further to improve this model.

License

SeaArt-Furry-XL-1.0 (and thus this derivative model) falls under Fair AI Public License 1.0-SD license, which is compatible with Stable Diffusion models’ license. Key points:

1. Modification Sharing: If you modify SeaArt-Furry-XL-1.0 (and thus this derivative model), you must share both your changes and the original license.

2. Source Code Accessibility: If your modified version is network-accessible, provide a way (like a download link) for others to get the source code. This applies to derived models too.

3. Distribution Terms: Any distribution must be under this license or another with similar rules.

4. Compliance: Non-compliance must be fixed within 30 days to avoid license termination, emphasizing transparency and adherence to open-source values.

The choice of this license aims to keep SeaArt-Furry-XL-1.0 (and thus this derivative model) open and modifiable, aligning with open source community spirit. It protects contributors and users, encouraging a collaborative, ethical open-source community. This ensures the model not only benefits from communal input but also respects open-source development freedoms.

The training script I used can be found here:
https://github.com/p1atdev/LECO

if I need to do basically anything else to maintain compliance I will literally just delete this model because I honestly don't care enough to attempt to maintain some stupid legal compliance.

描述:

Trained with recommended prompt structure

Trained with tags "detailed_background, detailed_character, detailed"

训练词语: