Hi, just trained my OC lora, Araya（荒弥）。
I suggest base model such as Plant Milk - Flax, or Indigo Furry XL.
Please try Hires fix when using this lora to get best experience.
15/Dec/2025:
A ZIT version of lora for Araya. due to dataset, this lora is suggested to use for base images ,and render them with SDXL models.
I am not very good at writing prompts in natural languages, so maybe I don't have good understanding in ZIT models and hence this lora.
My experience: Good prompts get good images, but bad prompts won't bother me at all. I use ZIT and Araya_ZIT to generate base images, and use them to generate latent images , rendering them with specific denoise ratio (0.3~0.5 if I only want to transfer styles without changing any details; 0.7~0.9 if I only want to get color blocks transferred to SDXL models.)
This technique is very convenient to get magical compositions ! But it's not very friendly for Forge users...
??/Nov/2025:
Somehow Indigo Furry XL make Araya stronger, but this depends on your other loras. Please try it out.
I generated the pics with Plant Milk - Flax .
You can use it with ' araya,1boy,' , but there is an outfit you can trigger with 'black cloak'.( 'cape' in v1.0)
V2 : Best weight：0.5~1.
solved problems in V1
V1: not recommended
!!! known problem: higher weight will make the background more simpler and orange when generating DARK Environments. I will try to fix that these when I get more dark samples .