A model merge of my favorites models: animayume_v10BaseFinal and animaika_v44 with anima_aestheticV10.
Drop the model in comfyui to see the exact recipe
Prompting advice:
The model itself is very flexible and has no predominant style, it's designed to be used with @artist tag or a style lora.
The first 5 preview images are raw model outputs with no lora, detailer and edits. The other i used a turbo lora and a bunch of custom nodes, i wanted to create a separate post but they ended as previews.
recommended turbo lora: https://huggingface.co/sorryhyun/anima-turbo-4step with a strength of 0.45, cfg 1.8, sampler and scheduler of your choice and 10 or 6 steps (depending of the sampler)
Description
FAQ
Comments (2)
Why [score_7::0.3] in the prompt? is score 7 too strong?
It's a personal preference to obtain the original official art look i like.
For people who don't know: [score_7::0.3] is different from (score_7:0.3), by using https://github.com/asagi4/comfyui-prompt-control you make the tag score_7 effective only for the first 30% of the sampling