Trained on the work of Nononononoway. It pushes images toward a glossy 2.5D style and excels with arthropods, monsters, post-apocalyptic settings, and machines. It will often be able to push anime based character loras into a more 2.5D presentation as well without forcing too much of the dystopian properties.The Flux 1 version is best used with narrative style prompts. The Illustrious and Pony models work better with tags but should support a mixture of tags and natural language. The Chroma and Anima lora should work equally well with tags, narrative prompts, or a combination of both. Likewise the z-image and Flux 2 loras should work equally well with tags, narrative prompts, or both. Though narrative heavy with a few keywords tends to work especially well. Though with z-image prompts if you're using both then it's best to begin with the trigger phrase, then the natural language description, THEN the tags. The chroma and z-image lora also benefit more from a description of the art style you want. With something like "in the Nononononoway style with a combination of x and y that promotes z" or the like.<h1 id="prompting">Prompting</h1>Just using a tag of 86_(nononononoway) should be enough for this style with all of the loras. However, you can often strengthen it a bit with "86_(nononononoway), This nononononoway style digital illustration is done in a style showcasing detailed shading to evoke gothic urban horror." especially with the more narrative focused models like z-image.<h1 id="examples">Examples</h1>Scorpion girl: 86_(nononononoway), nonoscorpiongirl, A detailed anime-style illustration of nonoscorpiongirl a pale young woman with long flowing silver-white hair and grey eyes. Her face and neck show intricate crack-like markings on her humanlike exoskeleton resembling porcelain damage or scars. She wears a mechanical mouth restraint with visible metallic teeth elements across her lower face. She's dressed in a flowing white Victorian-style dress with corset lacing at the chest. Her dark amber colored scorpion tail can be seen behind her extending from her lower back with her tail's stinger raised to waist level. Rendered in the nononononoway style the artwork blends hyper-detailed almost photorealistic textures with the clean expressive lines of anime character design creating a surreal high-contrast scene of ethereal dark fantasy paired with dramatic cinematic lighting. , braces, cracked cheeks, cracks on neck, 1girl, arthropod_girl, cracks on skin, cross-laced_clothes, cross-laced_dress, dress, face cracks, girl made of porcelain, grey_eyes, half-closed_eyes, long_hair, ringed_eyes, scorpion_girl, scorpion_tail, tail, white skin, white_dress, white_hairMoth girl: 86_(nononononoway), mothgirlnono, The pale-skinned figure of Mothgirlnono stands with long flowing silver-blue hair and piercing black eyes. Handcuffs dangle from her wrists. Multiple pairs of enormous iridescent butterfly wings extend from her back. Her butterfly wings are detailed with intricate veining and shimmering with holographic hues. She wears a dark blue short sleeved blue shirt paired with blue cargo pants. , 1girl, antennae, arthropod_girl, barefoot, blue_eyes, blue_hair, butterfly_wings, cuffs, expressionless, handcuffs, insect_wings, long_hair, moth_girl, multiple_wings, pants, cargo_pants, blue_pants, short_sleeves, solo, wings<h1 id="versioning">Versioning</h1>The 1.5 branch is an experiment with a fraction of synthetic data on top of the prior dataset used with 1.0. I'm still not really sure what I think of it in comparison. Though it did provide a few new concepts (Tsumugi Shirogane from Danganronpa V3 is more or less promptable from the 1.5 z-image lora for example).

Nononononoway Style

This Chroma lora should work with both tags and narrative descriptions for the prompts. Though in general I find that the art style comes out well with this: "A detailed anime-style illustration of INSERT SUBJECT OF IMAGE. Art style: high-quality anime illustration, detailed shading, gothic horror aesthetic, professional digital art quality, 4K resolution."This prompt can also be good to add some extra detail: (aesthetic 25:1.5), (aesthetic 11:1.5), (aesthetic 10:1.5), (aesthetic 9:1.5) ,highly detailed, intricate details, fine details, sharp focus, crisp edges, high definition, textured skin, "realistic skin texture, detailed facial features, masterpiece,best quality, ultra detailed, 4K, 8K, high resolution, professional digital art, studio quality, photorealistic details,detailed background, intricate environment, atmospheric lighting, detailed shadows, complex scene composition, detailed skin texture, detailed facial anatomy, A detailed anime-style illustration of INSERT PROMPT SUBJECT. Art style: high-quality anime illustration, detailed shading, gothic horror aesthetic, professional digital art quality.Check out the gallery for examples, setps, samplers, etc as I'm still unsure enough of Chroma to say much definitively about what works and what doesn't. I am a little concerned that I might have undertrained this lora. If I feel that way after some time has passed I'll probably try continuing the training for a while.While the whole horror-tech and insect apocalyptic vibe was my main goal, the Chroma version seems to do quite well as a standard anime style.

Files with SHA256: 9bfa2bf2ea436b76ad078549cb04daca4f4f106d58659bfcf5364d6beb0e20f6