[Style filter] Xu Er thick paint composition light texture enhancement - Illustrious
推薦提示詞
XUER guangying, 1girl, solo, long hair, breasts, looking at viewer, bangs, <lora:绪儿 厚涂构图光影质感增强V3:0.9>, (artist:quasarcake:0.8), extreme aesthetic, (wlop:0.6), honjou honjou raita, lack, rella, raita, wanke, good quality, year 2024, year 2023, Visual impact, A shot with tension, ultra-high resolution, 32K UHD, sharp focus, best-quality, masterpiece, Emotionalization, unconventional supreme masterpiece, masterful details, newest, very aesthetic, temperate atmosphere, absurdres, with a high-end texture, in the style of fashion photography, best quality, (Visual impact:1.2), giving the poster a dynamic and visually striking appearance, impactful picture, offcial art, colorful, movie perspective, amazing quality
推薦反向提示詞
(worst quality, low quality, very displeasing, lowres), (interlocked fingers, badly drawn hands and fingers, anatomically incorrect hands), blurry, multiple views, comic, censored, jpeg artifacts, watermark, patreon logo, patreon username, web address, text, logo, artist name, safe_neg, signature, unfinished, displeasing, oldest, early, chromatic aberration, artistic error, scan
推薦參數
samplers
steps
cfg
clip skip
resolution
other models
推薦高解析度參數
upscaler
upscale
steps
denoising strength
提示
Recommended LORA weight: 0.8-0.9
Recommended HD redraw rate: 0.42
Different samplers provide varied painting styles: DPM++ 2M for thick coating texture, Eulera series for 2.5D and anime style, DDIM CFG++ for hand-painted style
版本亮點
这是个风格lora.主要提供画风,减少AI感。
训练集用了更复杂的构图,更精细的场景。
和一些比较难写出来的动作。
例图用大模型: https://www.shakker.ai/modelinfo/465ea8f1653a4a00bbe98609129f9c3f/LORA?from=personal_page
推荐LORA权重::0.8-0.9
推荐高清重绘率:0.42
在不同的采样器下这个lora提供不同的画风。
DPM++ 2M 系列厚涂质感更好。
Eulera 系列画风更清晰会偏向2.5D和动漫风。
DDIM CFG++ 系列画风更偏手绘感。
触发词: XUER guangying,
提示词参考例图和返图区。(点击最新查看V3例图)
都是直出图大家放心复刻。
創作者贊助
Case diagram with large models: https://www.shakker.ai/modelinfo/465ea8f1653a4a00bbe98609129f9c3f/LORA?from=personal_page
This is a style lora. It mainly provides painting style and reduces AI sense.
The training set uses more complex compositions, more elaborate scenes.
And some of the more difficult actions to write.
Case diagram with large models: https://www.shakker.ai/modelinfo/465ea8f1653a4a00bbe98609129f9c3f/LORA?from=personal_page
Recommended LORA weight: : 0.8-0.9
Recommended HD redraw rate: 0.42
This lora provides different painting styles under different samplers.
DPM++ 2M series thick coating texture is better.
The Eulera series has a clearer style, leaning towards 2.5D and anime.
DDIM CFG++ series painting style is more hand-painted.
Trigger word: XUER guangying,
Prompt word reference example diagram and backmap area. (Click to view the latest V3 example)
It's all straight out so you can copy it.
这是个风格lora.主要提供画风,减少AI感。
训练集用了更复杂的构图,更精细的场景。
和一些比较难写出来的动作。
在不同的采样器下这个lora提供不同的画风。
DPM++ 2M系列厚涂质感更好。
Eulera系列画风更清晰会偏向2.5D和动漫风。
DDIM CFG++系列画风更偏手绘感。
触发词:XUER guangying,
提示词参考例图和返图区。都是直出图大家放心复刻。