CivArchive
    6suan Styles - Anima - Concept Art
    NSFW
    Preview 121825625
    Preview 121825880
    Preview 121825525
    Preview 121825631
    Preview 121825672
    Preview 121825703
    Preview 121825705
    Preview 121825744
    Preview 121825787
    Preview 121825792
    Preview 121825812
    Preview 121825833
    Preview 121825845
    Preview 121825876
    Preview 121825892
    Preview 121825916
    Preview 121825918
    Preview 121826213
    Preview 121825950
    Preview 121826434

    Collection of LoRAs based on different styles from 6suan for different models (for now, Pony and Anima)

    Updated with Concept Art Anima LoRA.

    Updated with Main Style for Anima LoRA. Also new Main Style using new dataset updated with enriched dataset with new artworks.

    Training LoRAs for Anima was an absolute breeze. Using too. Finally no more tag-list soup (though because I'm lazy - old prompts weren't converted).

    Since it's now also includes Anima, I changed a requirments for it a bit. Do not sell derivatives of this LoRA, I don't encourage donating Buzz or any other currency for use of this model, it was created for research purposes and love of the game and not for monetary gains. (or somehow like this Anima license was explained to me). You can do whatever you want with outputs, IDGAF. Have fun.

    All Anima showcase images were generated using ComfyUI, so workflow embedded inside. I tried using as less custom nodes as possible. For Concept Art I recommend using just simple Upscale using Remacri and then optional FaceDetailer (UltimateSDUpscale smoothes out the cell-shading concept art style). For Main Style - UltimateSDUpscale + FaceDetailer is the best combo. (Unfortunately Anima - can't work with Latent Upscale, so UltimateSDUpscale with low denoising is the only option, or straight up using simple Upscale by Model node).

    During generation of showcase images, found some minor problems with Anima. Like... it doesn't have that much of different species understanding as Pony. Or dog breeds (it all dog for it). But for 2B model that capable of text rendering and natural language understanding - it's rather a small problem. Anima is still the greatest newest shiny thing out there (Z-Image was kinda dissapointing).

    This LoRA is a derivative of a CircleStone Model and has been modified from the original. It is subject to the CircleStone Labs Non-Commercial License.

    "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."

    Description

    I'm absolutely in love with Anima. Can't wait for full model. Training this was easy as hell, tagging (or more like writing two sentences in natural language) same - VLMs really good at this.

    This is for Anima-Preview-2B, not SDXL, not Pony, not Illustruous, not Noob or whatever SDXL based new gooning checkpoint is out. Only for Anima.

    It based on more concept art like arts from 6suan. Main style LoRA ready, but now I'm busy making showcase images, probably gonna release it tomorrow.

    FAQ

    LORA
    Anima

    Details

    Downloads
    55
    Platform
    CivitAI
    Platform Status
    Available
    Created
    2/21/2026
    Updated
    4/27/2026
    Deleted
    -
    Trigger Words:
    @6suanconcept

    Files

    6suanconcept_anima.safetensors

    Mirrors