CivArchive
    WAM! - V1
    NSFW
    Preview 9635833
    Preview 9635840
    Preview 9635846
    Preview 9635844
    Preview 9635843
    Preview 9635847
    Preview 9635848
    Preview 9635849
    Preview 9635850
    Preview 9635851
    Preview 9647530
    Preview 9647552
    Preview 9647553
    Preview 9647538
    Preview 9647548
    Preview 9692420
    Preview 9717188
    Preview 9720135
    Preview 9726360
    Preview 9728908

    It's WAM! (Wonder Anime Mix)

    An Anime Model with clear style, beautiful and detailed characters and background.

    it got a large range of clothings, gesture, posture, action and body shape.

    NSFW and H compliant.


    Prompting in a WAMazing way

    First thing first, i used (masterpiece, best quality, highres) at the beginning of the prompt, and (worst quality, low quality:1.4), (bad hands, bad fingers, deformed hands, deformed fingers, malformed hands, malformed fingers, fewer digits, wrong hand:1.2) for the Negative Prompt.

    I didn't test TI/Negative Embeddings because, well, i don't like them so :p

    Now, this model can produce amazing thing but it can be tricky to use.

    A simple short prompt can sometime produce more better things than an over complicated long prompt.

    Also, using too much quality or aesthetic tags can result in something crappy, while less will produce somethings beautiful. So be carefull with your use of thoses tags.

    But the most tricky part are the view tags (face, upper body, cowboy shot, from below, from above, from behind etc...).

    Sometime they won't work. More exactly if, for example, you use shirt, skirt, pantyhose, boots tags and want to use upper body.... well this one will, most likely, transform into a cowboy shot and maybe even further view.

    So, like i said earlier, be aware of your prompt and simple one are sometime way better.

    As for LoRa, i got few that didn't work very well (mostly characters one), but you shouldn't have trouble with most of them.



    Generating in a WAMderfull way

    I used the following settings for all the image of the showcase :

    • Sampler : DPM++ 2M Karras

    • Sampling Steps : 40

    • CFG : 7

    • HIres Fix :

      • Hires Steps : 10

      • Denoising strength : 0.3

      • Upscaler : 4x-UltraSharp

      • Upscale : x2

    I got the best result with thoses settings on local with my old PC (a 6 years old 1050TI 4gb as GPU ^^;).

    So, if you use a more powerfull computer (or Google Collabs or other notebook), i encourage you to try more demanding settings and see the output :)



    The UltiWAM Recipe

    All the models used for the merge can be found in the following public collection : https://civarchive.com/collections/964613

    Don't hesitate to try thoses models and and thumb up if you like them ^^


    I hope you'll enjoy it and have fun with it.

    Oh, and yeah you can produce an handsome looking guy wearing a pink sweater with it :D

    Description

    Initial Release.

    Read Model Details for more info.

    FAQ

    Checkpoint
    SD 1.5

    Details

    Downloads
    556
    Platform
    CivitAI
    Platform Status
    Available
    Created
    4/11/2024
    Updated
    4/22/2026
    Deleted
    -

    Files

    wam_v1.safetensors

    Mirrors

    CivitAI (1 mirrors)
    Other Platforms (TensorArt, SeaArt, etc.) (1 mirrors)

    Available On (1 platform)

    Same model published on other platforms. May have additional downloads or version variants.