Please take a quick glance at the following notes before use. This model is relatively complex as it involves interactions between background crowds, spatial elements, and foreground girls.模型比较复杂，涉及背景人群、空间以及前景女孩的互动，使用前最好速读下面的简介。

V1 = 10 epochs | V1-alt = 15 epochs(Slightly Overfitting) + 10 epochs overlay training with (clear heels + new scene) images

V1 = 10 epochs | V1-alt = 15epochs(有些过拟合) + 10epochs叠加训练 (透明高跟+新场景图片)

V1 Info (Transparent Platform Heels)

model strength < 0.8 is generally recommended; otherwise, you might encounter body merging issues.

Scene Prompts

• DJ Booth + Dance Floor dance floor,dj booth,nc_scene

• Bar Area bar area,bar counter,nc_scene

• Dance Floor square stage dance floor,stage,nc_scene

• Sofa Area lounge area, sofa area,white sofa,black side table,nc_scene,moody atmosphere,nc_scene

• Elevated Platform nightclub,nigtclub's stage,stage, elevated square platform, (very high square platform:1.3),peole in background gathered ((below the elevated platform):1.4), surrounding it and ((looking at girl):1.4), at the base of the stage.nc_scene

Elevated Platform 拟合程度高度不足，需要大量的 Prompt 来从 FLUX 本身的生成画面中汲取大量的视觉元素再润色，本身 LORA 的效果只占了一小部分

Above are the scene types,Check Background Reference for more style

Action Prompts

• Dance dancing,(seductive pose:1.2), (sensual dance:1.3), (provocative:1.1), dynamic pose, energetic, (arms outstretched:1.2), hands reaching out,

• Half-Squat standing,half-squat, (bending knees, knees bent, spread legs,low stance, dynamic pose,leaning forward):0.6, (legs bent:0.7),hands on knee,,

• Deep-Squat squatting, full-squat, deep squat, low squat, ass to ground, sitting on heels, (buttocks below knees:1.2), (very low:1.1), dynamic pose,

• Single Knee kneeling, single knee kneeling, hand on knee, one knee down, (one knee on ground:1.2), (leg bent:1.1), dynamic pose

• Full Knee kneeling,( full kneeling, both knees down):1.3, (knees on ground:1.2), (sitting on calves:1.1), hand on thigh, static pose

• Pole Dance (pole dancing:1.3), (on pole:1.2), (dance pole:1.1), flexible,graceful,holding pole rod

• Posing photo posing,looking at viewer, sultry expression

• celebrating_liquor celebrating, cheering, holding liquor bottle, holding overhead

• posing_photos_with_liquor_bottle posing, group photo, looking at viewer, sultry expression, holding liquor bottle

Cloth Prompts

Only 14 images,Low success rate for some clothes

14 张图片根本不足以学习一些复杂衣服的细节，比如心形比基尼的细细的系带。而 Gem Bikini 大概率是因为 Girl-in-scene 数据集图片中有类似的银色链条+钻石的 rhinestone bikini，它能得到一定程度的有效学习和复现，但是很容易复现为 Girl-in-scene中的 rhinestone bikini 比基尼样式。至于款式较为简单的服装，模型更多的是从 GIrl-in-scene 中图片中去学习相应的服装，而从这个款式的一两张图片中学习到的特征很少。

• Red cutout dress (red cutout dress:1.3), (red cutout mini dress:1.2), (one-shoulder dress:1.1), (choker:1.0),

• Gem Bikini (gemstones with chains bikini:1.4), (blue gems:1.2), (minimal clothing:1.2), revealing outfit

• Gold Bodysuit (gold bodysuit:1.4), (shiny clothing:1.3), (latex clothing:1.2), (lace stockings:1.3), (white stockings:1.2), (garter straps:1.1),

• Neon green (neon green clothing:1.4), (fluorescent green:1.3), (long sleeve crop top:1.2), (bikini bottom:1.1),underboob,

• Fishnet bodysuit (fishnet bodysuit:1.4), (mesh clothing:1.3), (cutout bodysuit:1.2), (black clothing:1.1),

• Sparkling dress (sparkling dress:1.4), (rhinestone dress:1.3), (cutout dress:1.6), (one-shoulder dress:1.2), (black dress:1.1),

• White crop top (white crop top:1.3), (green trim:1.2), (bikini bottom:1.1), (underwear:1.1), round neck,short sleeves,tight,

• White mini dress (white dress:1.4), (mini dress:1.3), (bodycon dress:1.2), (shiny clothing:1.1), (silk dress:1.1), (satin dress:1.1), spaghetti strap, frilled trim, choker, white gloves,

• Rhinestone bikini (bejeweled bikini:1.4), (rhinestone bikini:1.3), (sparkling clothing:1.2),

• Black bodysuit (black bodysuit:1.3), (shiny clothing:1.2), (latex clothing:1.1),

• Heartshape bikini(not work) (heart-shaped bikini:1.5), (red heart:1.4), (pasties:1.3), (string bikini:1.2), (minimal clothing:1.1),

因为没有图片数据支撑，两双靴子只能从 girl-in-scene 图片中有限学习，同时大量融合了有图片数据支撑的 Clear Platform 的特点。

• Clear Platform wearing nc_heels,clear platform high heels with clear straps, the heels' straps are visible,

• Ankle Boots wearing ankle boots

• Knee-high Boots wearing knee-high boots

• Shoes not visible shoes not visible

Multi-Girl Control

For more advanced control over individual girl's poses and orientations, consider trying techniques like Regional Prompting.

3girls total in foreground,  (1girl from front:1.4),2girls from behind,

-

 3girls total in foreground, 2girls squatting, 1girl standing,

Special Scene Prompts: Diagonal Performing Stage

Dynamic shot,looking at performing on a ((diagonal stage edge):1.4) that extends across the frame. The audience is visible on the left side of the frame, below the stage, while the girl is prominently featured on the right side, above the stage.nc_scene,

Details here

A girl on stage, positioned in the upper left of the frame, with the stage floor creating a sharp diagonal line from top-right to bottom-left across the image. Below and to the right, the audience is visible, looking up.nc_scene,

Details here

Training Parameters:

147 girl-in-scene images | 14 clothing images | 21 heels images | 15 hands images | 53 pure scene images

_250 images, batch size 4, 10 epochs, 5 repeats, network_dim 32, fp16 ⚠️_train_t5xxl = false network_train_unet_only = true

---

V1 Alt Info (Clear Heels)

• convert Transparent Platform Heels to Clear Heels.

• thinner girls

• add 1 new scene type via caption refining.

New Scene Prompts

通过标注修改，强化了 elevated platform 概念的学习，不用像 V1 那样使用大量 prompts 描述借助 FLUX 基础模型来绘制这个平台了。

• elevated platform next to a DJ booth nightclub,nigtclub's stage,DJ booth,(on a very high elevated platform next to a DJ booth, crowd below platform):1.2,nc_scene

Training Parameters:

Overlay Training

_94images, batch size 4, 10 epochs, 5 repeats, network_dim 32, fp16. ⚠️_train_t5xxl = false network_train_unet_only = true

+

_250 images, batch size 4, 15 epochs, 5 repeats, network_dim 32, fp16.⚠️_train_t5xxl = false network_train_unet_only = true

Beta Info

Clear Heels are embedded

Training Parameters:

Overlay Training

118+34 girl-in-scene images | 21->16 heels images | 37+11 pure scene images

61images, batch size 4, 10 epochs, 5 repeats, network_dim 32, fp16.

+

_176 images, batch size 3(09 2,918 4), 18 epochs, 5 repeats, network_dim 32, fp16. ⚠️_train_t5xxl = false network_train_unet_only = true