Chroma Lora Version 3: I have already created the lora, it is with tensor.art training, well I should have preferred to train with diffusion-pipe but if there is an error code in the console cause of the lora, ignore it. Otherwise it works great. 👍

pony diffusion v6 xl was successful for lora training.

All pony diffusion models can generate on:

(Pony) normal models: ✅️

Anime models: ✅️

Cartoons like (ToonYou or toonify) models: ✅️

Photorealistic models: ✅️

Animatediff: ☑️

for photorealistic models you need:

extra models lora:

https://civarchive.com/models/310964/furry-enhancer-sdxl

https://civarchive.com/models/228974/great-lighting-slidersntcaixyz

Vae models:

https://huggingface.co/Mistermango24/fp16-fix-vae-sdxl/tree/main

embedding:

https://civarchive.com/models/148131?modelVersionId=165259

https://civarchive.com/models/148131?modelVersionId=165424

***[(only ziprealism_neg and ziprealism)]***

how is that supposed to works?

steps: 60

cfg scale: 7

clip skip: 2

HyperTile: True

Furry Enhancer (Lora) weight: 0.8-1.0

Great Lighting (Lora) weight: 1.0

any character (Lora) weight: (0.2 - 0.8) but you can set it to (0.4 - 0.6), if you want.