This is a style lora trained on images by artist Marmaladica. I also trained a <a target="_blank" rel="ugc" href="https://civarchive.com/models/2053408/marmaladica-artist-style-locon-with-dora-weight-decomposition">LoCon version</a> for illustrious, which seems to work better on checkpoints with a strong default style. If you don't use such a checkpoint this standard lora version should be fine. I used the same prompts and seeds for the showcase pics for both illustrious versions, so you can compare the outcome. You can find some works of the artist <a target="_blank" rel="ugc" href="https://www.deviantart.com/marmaladica">here</a>, so see for yourself. You can even buy their art as <a target="_blank" rel="ugc" href="https://www.inprnt.com/gallery/marmaladica/">print</a>. There is no trigger word required for this one.Feel free to comment, if you have something to say about the lora and share your generated images on the lora model page.If you like the lora please have a look at the showcase pics and leave a reaction.

Marmaladica artist style - standard Lora [ANIMA/IL]

I tried different schedulers for the showcase pics of the anima version and choose the images that looked good to me. Used mostly the same prompts as for the illustrious version. I used anima-preview1 for generating all but one image, however some images were upscaled using img2img with CAT Anima, since it seems to work slightly better with MultiDiffusion img2img. The style seems to work mostly good with anima-preview at weight 1.0, but if you want to use it with CAT Anima or another checkpoint, you should increase the weight. For more tips using anima look at the description of my kyhustyle anima LoRA!

Files with SHA256: 9808fed3b154116488f38a3ded3c06b2ba179d6f04d3655e8c5a0eaf38cd5dc6