Sign In

Chaos Warrior | Conan Exiles Legacy | Anima

0

Download

1 variant available

bf16 SafeTensor

ChaosWarrior_Anima.safetensors

BF16, good balance • 87.6 MB

Verified:

Type
LoRA
Stats

25

0

Reviews
Published

Sep 10, 2026

Base Model

Anima

Training
Steps: 1,000
Epochs: 10
Usage Tips
Strength: 0.8
Hash
AutoV2
F338D33301
Trigger Words
chaos_warrior_cl, horned helmet, full face mask, glowing eyes, fur cloak, chainmail, dark armor, skull belt buckle, chain belt, gauntlets, tall boots
chaos_warrior_cl, 1girl, solo, missing headwear, short hair, white hair, black eyes, black eyeshadow, fur cloak, chainmail, dark armor, skull belt buckle, chain belt, gauntlets, tall boots
default creator card background decoration
Reactions - 9244

9.2K

Followers - 389

389

Likes - 858

858

Acolyte Badge

License:

Anima

The Anima Model is licensed by CircleStone Labs LLC. 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

00138-- [832x1216] -- [3724998251] -- [None] -- [anima_baseV10].png

If you like my work and would like to leave a review, go for it. Motivation is motivation. I do this for fun mainly.


This LoRA is part of a Collection!


Model Information

A Clothing LoRA from Conan Exiles Legacy. The armor set is called Chaos Warrior, it came from a mod called Lexa's Exiled Lands Improved.

It's an outdated mod and it pisses me off because I wanted to cover Ancient King from the mod but it has buggy textures.

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


Helpful Information

Sample images were not meticulously hand-crafted one by one in Forge Neo and may not be the best representation of what the LoRA can do. I used semi-decent wildcards, and generated several batches of images across multiple checkpoints and picked the best 20.

Checkpoints I typically use


Metadata is not scrubbed so feel free to reference the sample images for whatever you please.


Training Notes

Every outfit is trained at 1000 steps (50 images, 10 epochs, 6 repeats, 3 batch size). They all get 1000 steps to look as good as they can. I'm an experimental person, so per outfit target steps wasn't interesting to me. You get the 10th epoch.

Every single outfit's dataset consists of in-game screenshots in an outdoor day-frozen area originally ideal for a base, but instead ideal for screen archery. Standing poses with pose/angle/framing variations (include multi angle frozen walk/run animation), Devious Desires and Tot Custom were used.

Every outfit is female only, not because I'm a degenerate (I am) but because it's too much work to support male versions and I think they're ugly.

Load my safetensor into LoRA Metadata Viewer if you're curious about any settings. I don't use nasty stuff like AI Toolkit so the metadata should be there.


Prompt Usage

Here is everything you should know this LoRA. Take note this is only for LoRA-specific information, the use of expressions, poses, angles, framing, background and the sort is your job.


Chaos Warrior (in-game preview)

chaos_warrior_cl, horned helmet, full face mask, glowing eyes, fur cloak, chainmail, dark armor, skull belt buckle, chain belt, gauntlets, tall boots


Other Tags: missing headwear, short hair, white hair, black eyes, black eyeshadow, outdoor, grass, wildflowers, forest background, daytime, 3d, game screenshot, realistic

  • You might generate my character if you prompt specific face details per-outfit.

  • If you want to remove a helmet, use missing headgear instead of the respective helmet tag. This tag is only headgear that cover the head, it may not work for circlets, earrings, antlers, etc.

  • I recommend putting 3d, realistic, game screenshot in your negatives, honestly. Especially if you're going for anime as every image is tagged with those.


Final Notes

Slapping this at the very end of my description(s) in an attempt not to annoy people as much.

1) I create content entirety for fun. I refuse to do this for money or to ask for money. I don't do commissions or anything of the sort. You can request things, but if I don't like it or care for it, you won't get a response.

2) If you are curious how I made something, just ask me. I don't mind sharing a dataset, training settings and the sort. I am a curious person. I'm not overly protective of my work and I am not worried about who does or does not jack off to it.