"My main motivation for creating these models was to produce more artistic and stylized furry art."
Check out my DeviantArt page and follow me: https://www.deviantart.com/furryartist2
These models were made with Yiff-E18, Zack3D (v1), BB95 Furry Mix, Fluffrock, Loras, and some others.
Documentation:
Main orientation (bias): "Male focus"
FurryArt_noise-offset: (This is my very first model!)
Note: This is a noise-offset model. This allows to produce high dynamic range images, you can use optionally CFG_Rescale_webui extensions.
Initial model: Yiff e-18
Standart Resolution: 512 p.
Resources: Works very well with "upscale method" or "donwscale method" like "Kohya hires fix" or "jank HiDiffusion"
FurryArt_B (Finished model):
Note: Similar to the previous model, but this is a no noise-offest model. Is more accurate for furry artists.
Initial model: "Noise-offset"
Standart Resolution: 512 or 768 p.
Works very well with "upscale method" or "donwscale method" like "Kohya hires fix" or "jank HiDiffusion"
FurryArt_Bv2 (Finished model):
Initial model: "B"
Standart Resolution: 768 p.
Works very well with "upscale method" or "donwscale method" like "Kohya hires fix" or "jank HiDiffusion"
FurryArt_C (Finished model):
Initial model: "B2"
Standart Resolution: 768 p.
Works very well with "upscale method" or "donwscale method" like "Kohya hires fix" or "jank HiDiffusion"
F. Art_C ver.2 (Finished model)
Note: It is an improvement of the previous model (merged with FluffyRock), more photorealistic
Initial model: "C"
Standart Resolution: 1024 p.
Works very well with "upscale method" or "donwscale method" like "Kohya hires fix" or "jank HiDiffusion"
FurryArt_FR (Discontinued model):
Note: This was first FluffyRock model merged. The purpose was to create a model more easy and flexible to write prompts.
Initial models: "FluffyRock <- C"
Standart Resolution: 1024p.
F. Art_FR ver.2 (Discontinued model):
Note: It is an improvement of previous model, and is more photorealistic.
FurryArt_D (Active model):
Note: This is a derivate of "FurryArt_B".
Initial model: "B2"
Standart Resolution: 768p.
FurryArt_R (Discontinued model, unstable):
Note: Model designed to be semi: photorealistic, realistic/3d,
also it can be unstable with LoRas, also a bit unstable with hands.
Initial models: "Fluffyrock <- B2"
Standart Resolution: 1024p.
FurryArt_R ver. 2 (Active model):
Note: express correction and inprovements for the model "R"
Anatomy stability
face stability
Lora stability
ControlNet stability
Initial models: "R and some others, but I don't remember.....😅😂"
Standart Resolution: 1024p.
FurryArt_R ver. 3 (Stable model, active)
Note: Improvement for R2 model. For this model I made improvement in sharpening and background, but I have seem this model have a kind of "human hands bug". I fixed the "human hands bug" in the new update.
Initial models: R2 + some "detail" Loras
Standart Resolution: 1024p.
FurryArt_GU SPECIAL EDITION (Stable model, active)
Note: Model derivate from FurryArt_B and Redwater, this is a very Artistic model, you can make smooth, painting images or very sharp, photorealistic images.
Note 2:
This project only was possible because of your help, thanks Gungie!
https://civitai.com/models/160293?modelVersionId=180346
Initial models: B <- Redwater (by "rebasin" merge method)
Standart Resolution: 1024p.
**Special models**
FurryArt R 4 (Stable model, active)
Note: improvement of the model "R3".
Standart Resolution: 1024p.
FurryArt LCM (Stable model, finished)
This is a Latent Consistency Model (LCM) merge. It is capable of producing very sharp images in just 3 or 4 steps with a CFG scale between 1 and 3.
Standart Resolution: 768p.
FurryArt_Hyper4 (Stable model, finished)
This is a merge model integrated into Hyper-SD. It is capable of producing sharp and detailed images starting at 4 steps with a CFG scale of 1 to 1.5.
Realistic: for realistic and semi-realistic images.
Artistic: for artistic and some semi-realistic images.
This is a merged model integrated into Hyper-SD. It is capable of producing sharp and detailed images starting at 4 steps with a CFG scale of 1 to 1.5.
Standart Resolution: 1024p.
FurryArt_XL (Stable model, active)
This is an "SD XL" merged model, finalized with a PCM LoRA on my modest computer (GTX 1060 6GB / 16GB RAM) using ComfyUI and the "mecha_merger" extension (a model merger focused on low memory footprint), and pruned with Automatic1111.
With this model, you can create consistent images in 8 steps using a normal CFG scale.
Standart Resolution: 1024p.
.....UNDER CONSTRUCTION......