IL v2.3 : New version, more flexibility. Better than IL v1
This version don't have a baked in VAE. I recommend to use SDXL VAE
Please consider supporting me so I can continue to make more models --> https://www.patreon.com/BB95FurryMix
Don't forget to join the Furry Diffusion discord server --> https://discord.gg/furrydiffusion
Since v3, this model uses e621 tags.
This model is a mix of various furry models.
It's doing well on generating photorealistic male and female anthro, SFW and NSFW.
I HIGHLY recommend to use Hires Fix to have better results
Below is an example prompt for the v7/v6/v5/v4/v3/v2.
Positive:
anthro (white wolf), male, adult, muscular, veiny muscles, shorts, tail, (realistic fur, detailed fur texture:1.2), detailed background, outside background, photorealistic, hyperrealistic, ultradetailed,
Negative:
I recommend to use deformityv6 and boring_e621_fluffyrock_v4, you can add bad-hands-v5 if you want
Settings 1.5 :
Steps: 30-150
Sampler: DPM++ 3M Karras, DDIM, UniPC or Euler A
CFG scale: 7-14
Size: from 512x512 to 750x750 (only v4/v5/v6/v7/v8/v9), up to 900x900 for v10/v11/v12
Denoising strength: 0.6
Clip skip: 1
Hires upscale: 2
Hires steps: 30-150
Hires upscaler: Latent (nearest)
Settings XL :
Steps: Below 30
Sampler: Euler A
CFG scale: 3-6
Size: 1024x1024
Denoising strength: 0.2
Clip skip: 1
Hires upscale: 2
Hires steps: Below 30
Hires upscaler: Remacri
Description
I think this is the best version I made. More realistic, better fur, better clothes, better NSFW
FAQ
Comments (19)
May I ask why I feel that v8 is not as good as v7 in generating realistic style pictures? I use exactly the same prompt, maybe the compatibility with lora is not so good?
Maybe ! Lora tends to crush realism
@Bumblebee95 After removing all lora, the result becomes normal, but I still hope that version 9.0 is more compatible with Lora, after all, the model itself still has limitations
@bananattrr933 isnt that a lora issue, some are more overbaked and totally change the artstyle way too much
@es69 I agree, I think v8 adds more fantasy elements to the training process. As tested, v8 performs better in fantasy and v7 performs better in sci-fi and modern
Has anyone tried to make a taker_pov drawing? Successfully, I mean 😉 if so... please, teach me your ways 🙏😅
taker pov don't have a lot of images in e621
Maybe with a lora ?
@Bumblebee95 yep, kind of a rare thing really.
Might try looking for one, thanks! On the other hand... I have too many Loras already 😂
@JxC Try this: https://rentry.org/trashcollects#img2img-examples
@eeeeeqeq wow! Thank you so much! Completely forgot about it 😯
Bumblebee95!
I love your model, great work. Best furry model I have tried so far. I wanted to tell you that I have just published a modification of it that generates more perverted buttholes because I'm a degenerate like that, and I have instructed people to give all credit and support to you. I hope this is okay.
Don't worry ^^
This is okay
excellent model, great work!
Thanks !
TFW you realize you've been using CLIP skip 2 and a VAE to generate images like for 48 hours TT_TT
ahah ^^
you must have got some strange results !
The MSE vae won't degrade the results, it will either not change it (because its already embedded) or improve it (because it wasn't embedded)