Resized to 70% of original (view original)
| Prompt | (masterpiece), (best quality), (beautiful hands), (detailed hands), ((ultra-detailed)), (expressive eyes, detailed eyes:1.3), blush, (light smile), (cute face), perfect lighting, colorful, vivid colors, (pink lipstick), solo, smile, (pale and pale skin:1.3), (beautiful medium boobs1.5), (round breasts:1.4), on back, bright backgrounds, night, detailed, (standing in the shade under the tree), a satisfied face, deep cleavage, (pale skin:1.2), 1girl, highres, intricate, collarbone, (cowboy shot), first-rate female bodies, perfect anatomy, tiara, (delicate skimpy outfit:1.3), ((random color hair)), ((random color eye)), <lora:coloured glaze-000018:0.8>, coloured glaze, looking at viewer, (arms behind back:1.2) |
|---|---|
| Negative prompt | (worst quality, low quality:1.4), monochrome, large head, extra digits, bad eye, text, cropped, jpeg artifacts, (blurry, blurry background:1.3), signature, watermark, (username), white bars, (interlocked fingers:1.3), (extra fingers, deformed hands, polydactyl:1.4), (bad anatomy:1.3), (extra limbs:1.35), censored, mosaic censoring, bar censor, censor, (loli:1.1), (loli face:1.2), (very long hair, long hair:1.5), (bad-hands-5:1.5), EasyNegativeV2, (nipple piercing:1.2), (dark skin:1.3), (pointy ears:1.2), EasyNegative, bhands-neg, bhands, (thighhighs, stocking:1.2) badhandv4, (wet:1.1), deepnegative, bad artist, bad _prompt_version2, (bare back:1.5), shiny, (partially visible vulva:1.4), (colored hair:1.3), nsfw, (bare legs:1.3), (swimsuit:1.4), (bikini:1.4), , EasyNegative, deepnegative |
| Sampler | DPM++ 2M Karras |
| Seed | 4162169055 |
| Steps | 30 |
| Cfg Scale | 7 |
| Model Hash | b6a65b07e6 |
- ‹ prev Search: night status:any status:any next ›
I`am use LoRA: Color "Glaze/Gorgeous Background".
LoRA on the website (civitai)
Short original description:
This is a style-based Lora, with an overall artistic style that is very elegant and beautiful, just like the example images. It can handle both characters and backgrounds, and there is a hint of manga-style due to some training data, although it primarily focuses on backgrounds.