This model was developed through model merging and fine-tuning with the goal of simultaneously producing results similar to the Illustrious model and the artistic style of the XL model.
It is an Anima model.
Based on an anime-style aesthetic, it supports a variety of art styles.
Description
FAQ
Comments (4)
Can we get just the u-net? My HDD is getting full...
This merge is degrade quality pretty significant, see example on my lora and base model
https://civitai.red/images/132251203
vs
this model https://civitai.red/images/132197464
It’s just a matter of personal preference.
I wasn’t satisfied with the base model, so I made this one myself.
I think you should use whatever produces images you like.
Fun checkpoint to use with img2img as it features good stability at .6 denoise. The clean lineart and stability comes at the cost of artist tags largely not working anymore though. Still recommended because the baked-in style is really nice!