Sign In

Lilith | Skyrim RaceMenu Preset | Anima

4

Recently loaded

Download

1 variant available

bf16 SafeTensor

Lilith_Anima.safetensors

BF16, good balance • 87.6 MB

Verified:

Type
LoRA
Stats

22

4

Reviews
Published

Nov 19, 2026

Base Model
Training
Steps: 1,000Epochs: 5
Usage Tips
Strength: 0.8
Hash
AutoV258449799A1
Trigger Words
masterpiece, best quality, absurdres, score_9, score_8, l1l1thch4r, 1girl, solo, black hair, medium hair, wavy hair, pink eyes, eyeshadow, lipstick

The CircleStone Model is licensed by CircleStone Labs LLC under the CircleStone Non-Commercial License. Copyright CircleStone Labs LLC. IN NO EVENT SHALL CIRCLESTONE LABS LLC BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH USE OF THIS MODEL.

Built on NVIDIA Cosmos

00558-[2071569490] -- [Lilith (Anima), Clothes] -- [oneObsessionAnima_v40] -- [832 x 1216].png

Information

RaceMenu Presets from Skyrim that have been selected to be turned into Character LoRAs. Some were made by me.

This is the Anima version. The Illustrious version is over here.


Prompt Usage

A Imperial RaceMenu preset I downloaded. Her name is Lilith. Decent looking enough.

Sample Prompt 1

l1l1thch4r, 1girl, solo, black hair, medium hair, wavy hair, pink eyes, eyeshadow, lipstick

Important Notes

- The trigger does not absorb the style. All images contain the tags 3d, realistic, game screenshot. The first two are important. Put them into negatives if you want an anime style.
- The clothes in the sample images are wildcards and do not represent the characters. The characters do not have any clothing associated with them, nor are they trained wearing clothing.
- Sample Images typically use anime style tags such as anime coloring or anime screencap, but you may see me use no tags to get the dataset look.


Disclaimer(s)

If you generate with my content, and you'd like to share it, please consider using the "Add Post" button at the bottom of the gallery.

If you don't like seeing your work being associated with AI as a mod author.. while I do understand, I do not offer my shoulder to you.


Training Parameters

For this particular set of outfits and characters, I used a Comfy workflow to remove the backgrounds and replace them, alternating with grey, white and black backgrounds. I added the images back into the dataset, increasing them up to 100 images.

This time around, characters are trained at 1000 steps (5 epochs, 4 repeats, 2 batch size) and outfits are trained at 1250 steps (5 epochs, 5 repeats, 2 batch size)