There's something I want to say
I don't have much experience in making lora models. I mostly make models for myself and publish them just for fun.
They are far from what I would call “very good” quality and I (most likely) make a lot of mistakes when tagging the dataset and ignoring trainer settings
Since I'm pretty limited on resources and have to use Civitai for training, so I'm limited in the number of models I can create. Unfortunately I can't make changes right away and see what happens because of these limitations
I'm still very happy that someone likes these models.
Base:
fiammetta, short hair, red hair, animal ears, red eyes
Standart outfit:
jacket, shirt, skirt, high-waist skirt, pheonix skirt, knee boots, necktie, id card, gloves, thigh strap
Divine Oath:
white shirt, red skirt, armored boots, shoulder armor, armor, cape
“Divine oath” had an extremely small set of pictures. Therefore, this skin may differ from the in-game version. I highly recommend experimenting with tags when generating images with this outfit
Description
Details
Downloads
263
Platform
CivitAI
Platform Status
Available
Created
8/29/2024
Updated
9/30/2025
Deleted
-
Trigger Words:
fiammetta, short hair, red hair, animal ears, red eyes
jacket, shirt, skirt, high-waist skirt, pheonix skirt, knee boots, necktie, id card, gloves, thigh strap
white shirt, red skirt, armored boots, shoulder armor, armor, cape
Files
Arknights_Fiammetta.safetensors
Mirrors
CivitAI (1 mirrors)
Other Platforms (TensorArt, SeaArt, etc.) (1 mirrors)
782498_training_data.zip
Mirrors
CivitAI (1 mirrors)