I suggest a lora strength around 0.8 along with quality tags to get similar results to the example images provided. Images were generated using Anima Base v1.
This lora is a blend from some of my favorite lora's from Illustrious into a new lora for Anima.
This lora utilized the following Lora's and checkpoint for training. I've included the training data for this if you would like to train your own.
Checkpoint
Lora's
Description
Image dataset remade reduced to 59 images from 100, captions improved, trained at 768x768 reduced from 1024x1024.




