(IL 0.1)--> IMG test soon

Usually for this I just want to test some tag-based and some caption-based training via this sites training.

All images are more than 1k resolution; some of those are tags with some caption from Google Gemini.

(Img collection from some anime screen movie)

some img on dataset usually have high contrast level.

most img i choose use golden ratio compostion, rules of third, and mix (I dont tags this)

main trigger word

cinematic anime universe 100%
cinematic 65%
hdwc styles 50%
other tags with caption

Example: A fiery anime-style character dominates the image, featuring vibrant red hair tied up with loose strands and striking red eyes that convey intensity, while dressed in a dark kimono adorned with flame-like patterns, holding a katana poised for action casting a dramatic and dangerous mood the backdrop is engulfed in darkness, contrasted by the radiant orange flames that seem to emanate from the sword illuminating the scene and enhancing the character's dynamic pose, sparks are scattered around her, suggesting intense energy and movement this composition highlights the contrast between light and shadow, accentuating the powerful, combative stance of the central figure.weapon, sword, red hair, red eyes, fire, holding weapon, hdwc styles, holding sword, solo, katana, gloves, ponytail, black gloves, looking at viewer, fingerless gloves, long hair, 1girl, flaming sword, 1boy, flaming weapon, japanese clothes, male focus, fighting stance, cinematic, epic scene

Search AI Models & Addons

very few results