This Lora is made to enhance to ability / quality to create anthro and furry, and creatures with sdxl. but you can also use it to enhance also animals or fur clothing, etc.
its trained and tested on my models but it work on also on other types of models include some pony models . i highly recommend to use it with the following models.
: Midgard Pony, Ratatoskr , Bifröst Project , Fenrisxl ,Yggdrasil, Midgard
the lora can be used for SFW and NSFW.
Changelog:
V1 This is the first Version and still in WIP
i recommend using a strength of 0.7-0.9
to get realistic output use the trigger word "anthro"
for more digital art style "furry"
V2.84
added around 200+ more training images (around 800 total) and did a complete retain
I recommend using the Model Ratatoskr or Bifröst Project
Please be so kind and keep the Showcase clean of heavy ???? Stuff and check out my other stuff too ;-)
and thanks to S1LV3RC01N for the help please check out her models too
V5.2
this version is completely new trained from ground up, over weeks. on the base of my Bifröst Project Model
better look
more supported models
more specimen
better in nsfw
etc
i will provide the complete wordlist in the future
can turn with a higher strength ( 1-1,3 ) basically any model into a furry model 👍😉
V6.1
Model got retained again with over 342K Steps and a Training time of 300+ Hours
More versatile
better details
better in NSFW
compatible with different types of model ( sdxl 1.0, lightning, hayper, turbo, pony)
the tagglist is inside the "training" folder i also included some example Pictures with comfy workflow embedded.
V1.0 Flux (Photoreal)
First Flux Version of the enhancer trained on parts of the SDXL Version of the enhancer but the tagging got optimized for flux. Its able to bring furries in nearly any Flux model.
note: it's "only" trained at 960 pictures yet.
Note: There will be 2 Versions: Photoreal, and Digital Art
the Digital Art is still in training and follow later.
it can do sfw and nsfw.
Have fun with this version of the Furry enhancer
I can recommend my Workflow for it.
Also please Check out my Flux Model Ratatoskr Flux
V3.420 [Flux] (Photoreal)
More detailed
several improvements
Trained on a way bigger 1.5K pictures database also changed also the tagging a bit too.
Note: the needed weight of the lora vary a lot between different models the base model is around 0.7 other like my Bifröst Project Flux more 0.3-0-4 for example
Flux ???? Mode V1
Some people had complained about the ???? part of the flux Lora this is my answer to that 😁😉
This Version is made for enhancing Flux models in the ???? Area but can also used in lower weight for enhancing overall
its trained on my Newdawn Flux Model .
Made for Enhancing Comic and Realistic output
NSFW in general ;-) Enhancing ???? Details
adding BJ % Sex
but in lower amount works as an regular enhancer too (max 0.6)
I recommend using the enhancer between 0.7 and 0.9 for Newdawn Flux
Works also with other models but keep in mind it was trained specific on this model so no one knows exactly how good it runs with other models ;-)
RealisticV1.3 (SDXL HEL)
This Version is trained and tested on my Model Ratatoskr V10.94 Helheim
its trained on the Furry Enhancer Flux Photorealistic database
note that this model is its a bit its own type of sdxl model that means its not working well with the sdxl base model or other sdxl types.
So if you use that stay with the Furry Enhancer 6.1
i have tested it with other models too it works also partly on pony models for example with my Midgard - Pony but also Hayden's YiffMix Realism did work so its likely that it works with other Pony and furry model too but no guarantee for that
Weights: For the Ratatoskr V10.94 Helheim I can recommend a weight between 0.3 and 0.8 Note: with a higher weight the model will lose creativity
Flux Krea V2.3
This Enhancer Version is trained on the Flux Krea Model
How the model works and technical data details visit https://huggingface.co/black-forest-labs/FLUX.1-Krea-dev (FP8)
This lora is recommended for realistic output on animals, anthro, furry and creature's
The model is tested with the FLUX.1-Krea-dev
and my Ratatoskr Krea V11.39 Model for settings:
For the FLUX.1-Krea-dev i had my best results with 0.85 strength
because part of it are already in my Ratatoskr Krea V11.39 the settings are between 0.2 and 0.7
This version is for Qwen Image
its trained on the furry enhancer database at 118700 steps
whats is better than with flux?
more coherent
better understanding overall
trained in natural language and e631
knows many specim
realistic but can also do drawings
made for sdw and nsfw
Furry Enhancer Qwen Image V1.3
For settings:
I choose a strength at 0.85 of the lora can recommend between 0.5 and 0.9
i used the following setting:
30 steps
Euler / Beta57
CFG 3.5
For Workflow:
you can download the picture and but in comfy or download the workflow found in the training files. or use the all in one workflow https://civitai.com/models/1986585?modelVersionId=2248776
Please give a like to support my work.
it's really not easy to make such lora also the costs are very high!
made by freek22
This disclaimer is not here without reasons. please have in mind this is like a virtual showroom and posting extreme stuff is like 💩 on my desk in real live 😉
made by Freek22
please give a like to support my work
