An attempt to create a realistic ANIMA model.
Version 2.0
Version 2.0 was primarily tuned for the "er_sde" sampler, steps 12-16.
Run: Euler/simple (Beta), CFG 1, Steps 8-16
Text encoder: https://huggingface.co/circlestone-labs/Anima/tree/main/split_files/text_encoders
VAE: https://huggingface.co/circlestone-labs/Anima/tree/main/split_files/vae
Description
This version turned out a bit distorted and the prompt adherence has slightly worsened..
run:
sampler: euler, sheduler: beta, steps: 12-16, CFG: 1
FAQ
Comments (13)
Why isn't anyone publishing the prompt?
All the prompts were taken from here: https://civitai.red/images
@toya_san was just going to say that. recycled prompts for a different model. that's common on here.
@TheSplendidSpinna I was just too lazy to invent my own prompts, so I borrowed them from the examples. ))
I really like your models, thank you for making and publishing them!
However, I find that non turbo models follow my prompts much better. Is there any chance you could please release non turbo version of the new versions as well? Thank you greatly in advance.
When you train this, are you using a mixture of natural language and booru tags, like Anima used for its training? Or are you only using booru tags?
I don't train the models; I specialize in merging them and perfecting them until they yield results that please me.
@toya_san which models did you use in the merge?
@SoyKaf I use all the realistic models that are uploaded, and I also use some LoRAs.