Illustrious SNAFU - Super Nova Anime-Furry Unified

https://civitai.com/models/1194136

SNAFU is a NoobAI based Illustrious merge of several models and LoRAs combined and tested with goal of making a flexible mix that has a good base style, level of detail, and a mix of anime and furry knowledge to allow swapping between the two on the fly. It can easily produce NSFW, SFW, and/or nude characters with "featureless" bodies.

As a merge, this model owes its existence to Crody (https://civitai.com/user/Crody) and tkio (https://civitai.com/user/tkio) for NovaFurry/NovaAnime and NTRMix XIII, respectively. This model also assumes the restrictions as stated by Crody as follows:

Rules

You cannot use the generated images for commercial use if it's not edited (or just turning it to black and white) You can share images without any restriction if you don't monetize it

Advertising this model to outside is always welcome

Recommended Settings

Steps: 25 to 40

CFG: 4 to 5

Sampler: DPM++ 2M SDE Karras

Hires: DPED, 20 steps, 0.4 denoising

Prompt Recommendations

Oddly, I've noticed that in testing that the model can do well without the classic Illustrious positive and negative prompts. However, this being an Illustrious model means that using them anyway will not harm your generations. So, here's what I generally recommend:

Positive: masterpiece, best quality, amazing quality, very aesthetic, absurdres, {your prompt}

Negative: worst quality, low quality, lowres, old, oldest, sketch, error, bad anatomy, bad hands, url, text, displeasing, watermark, ugly, distorted, censored, abstract, signature, {your negatives}

BREAK: If you use this model in A1111 or Forge (or any other UI that supports the BREAK keyword), I recommend using it. In general, I use it to separate concepts. Especially in long prompts.

Example:

{quality and composition prompt} BREAK

{location and setting prompt} BREAK

{character and posing prompt}

BREAK

{clothing and actions prompt}

This isn't explicitly required, but you might find that it helps with coherency in many models if used in a UI that supports the keyword. Experiment with it and see if it helps!