license: creativeml-openrail-m
library_name: diffusers
pipeline_tag: text-to-image
tags:
- stable-diffusion
- stable-diffusion-diffusers
- diffusers
- text-to-image
- WarriorMama777
- NovelAI
- AnythingV3.0
- hesw23168
Orange Remix
I did a blind test of all the models at https://huggingface.co/WarriorMama777/OrangeMixs/tree/main/Models , then I found the most adorable one was BloodOrangeMix, the best picture composer was ElyOrangeMix, and the most coherent model was EerieOrangeMix_night.
Then I made a 50/50 blend of the first two, creating OrangeRemixBeta, the most adorable and creative model but bad at coherence, and finally I added 15% of EerieOrangeMix_night to that, creating my favorite Orange model!
It was a surprise that none of the Abyss models made it to the blend, but here's the full recipe to recreate Orange Remix:
STEP 1:
Add Difference 0.3 Primary Model
EerieOrangeMix_base
Secondary Model
NovelAI animefull
Tertiary Model
NovelAI sfw
Merge Name
EerieOrangeMix_Night
STEP 2:
Add Difference 0.3 Primary Model
Elysium_Anime_V2
Secondary Model
NovelAI animefull
Tertiary Model
NovelAI sfw
Merge Name
tempmix
STEP 3:
Add Difference @ 1.0 Primary Model
tempmix
Secondary Model
Gape60
Tertiary Model
NovelAI animefull
Merge Name
ElyOrangeMix
STEP 4:
Add Difference @ 0.3 Primary Model
AnythingV3.0
Secondary Model
NovelAI animefull
Tertiary Model
NovelAI sfw
Merge Name
tempmix-2
STEP 5:
Add Difference @ 1.0 Primary Model
tempmix-2
Secondary Model
Gape60
Tertiary Model
NovelAI animefull
Merge Name
BloodOrangeMix
STEP 6:
Weighted Sum @ 0.5 Primary Model
BloodOrangeMix
Secondary Model
ElyOrangeMix
Merge Name
OrangeRemixBeta
STEP 7: Weighted Sum @ 0.85 Primary Model
EerieOrangeMix_Night
Secondary Model
OrangeRemixBeta
Merge Name
OrangeRemix
STEP 8: No Interpolation Primary Model
OrangeRemix
Bake in VAE
color101VAE_v1.pt
Merge Name
OrangeRemixCVAE
Or something like that, I didn't keep notes!