Loading
Have a ComfyUI workflow? Get a readable Workflow Doc in your browser.Make a workflow readable.Try it now ↗
Lin, the cat girl gunner from Breath of Fire: Dragon Quarter (PS2). She belongs to an organization called "Trinity", which opposes the "Ranger" faction that the main character Ryu originally is aligned with. Before...
Model source
Source description
Lin, the cat girl gunner from Breath of Fire: Dragon Quarter (PS2). She belongs to an organization called "Trinity", which opposes the "Ranger" faction that the main character Ryu originally is aligned with. Before long they join forces to protect a girl named Nina from the Rangers, who wish her harm for some reason.
Okay, another obscure character from an obscure entry to a non-obscure jrpg series. Surprisingly popular fan art-wise...
Usage notes for different versions:
Anima: Works on strength 1. Mostly the dataset is the same as in the illustrious version. She can be generated with the hood on or down (remove "cat hood" and replace with "hood down" and "animal ears" for the hoodless version). Natural language can help with details but is not necessary.
Illustrious: Works on strength 1. Due to few new images in the dataset the hood comes off! Well, sometimes... Try adding "hood, cat hood, animal hood, hooded" etc. to negative prompts while prompting for "hood down" and it might just work.
Sources
1 sourceModel artifacts
1 artifactLin_(Breath_of_Fire_Dragon_Quarter)Ani.safetensors
safetensors · BF16 · 16.6 MB · SHA-256 e0fe495ac4e4…99c4 · Civitai
DownloadSource excerpts
1 excerptSource context: 364 downloads · Type LORA · Base model Anima
PonyXL: Works on strength 1. As in the SD 1.5 version, the hood stays on no matter what, if you want it off you'll have to do a lot of creative inpainting.
SD 1.5: Use strengths between 0.8 - 1 for best accuracy. As the dataset contained some images from in-game models, it has a slight tendency to make 2.5d-like images. If the character gets generated looking younger than her age of 21 years, you can add "adult female" tag or something similar. Also, the tail can be tricky to get right and the dataset contained no images of her without the hood on, so it stays on.
Hope you like the lora! :)
