🌟 Flux2Klein9B-ColorLight LoRA (v1)

色块驱动 · 任意角度打光 · 智能光影映射
Color-Block Driven · Arbitrary Angle Lighting · Intelligent Shadow Mapping

🔍 简介 / Introduction

Flux2Klein9B-ColorLight 是一款基于 Flux2Klein9B 架构微调的创新型文生图/图生图 LoRA 模型。它打破了传统提示词对光源描述的模糊性，实现了**“根据色块颜色与位置控制打光”**的核心功能。用户只需在画面相应位置涂抹对应颜色的色块（Mask），模型即可自动解析并生成符合物理逻辑的光影效果。无论是模拟夕阳侧逆光、霓虹灯补光还是局部聚光灯，都能实现“所见即所得”的交互式布光体验。

Flux2Klein9B-ColorLight is an innovative Text-to-Image/Image-to-Image LoRA model fine-tuned on the Flux2Klein9B architecture. It breaks the ambiguity of traditional prompt-based lighting descriptions by enabling "lighting control via color blocks and positions." Users simply need to paint color blocks (Masks) in specific areas, and the model will automatically parse them to generate physically logical lighting effects. Whether simulating sunset rim light, neon fill light, or spotlights, it delivers a "what you see is what you get" interactive lighting experience.

✨ 核心特性 / Core Features

✅ 色块即光源 (Color Blocks as Light Sources)：直接将画布上的颜色区域识别为光源发射点。红色色块生成暖红光，蓝色色块生成冷蓝光，白色色块生成自然补光。
Directly identifies color regions on the canvas as light emission points. Red blocks generate warm red light, blue blocks generate cool blue light, and white blocks generate natural fill light.

✅ 角度精准映射 (Precise Angle Mapping)：根据色块在画面中的相对位置（左上、右下、正前方等），自动计算光线入射角与阴影投射方向，实现真实的立体感塑造。
Automatically calculates light incidence angles and shadow projection directions based on the relative position of the color block (top-left, bottom-right, front, etc.), creating realistic three-dimensional depth.

✅ 多光源混合 (Multi-Light Mixing)：支持同时使用多个不同颜色的色块进行布光，模型能智能处理多光源叠加、色彩混合与主次光比关系。
Supports using multiple color blocks simultaneously for lighting. The model intelligently handles multi-light superposition, color mixing, and the ratio between key and fill lights.

✅ 实时反馈 (Real-time Feedback)：配合 ControlNet 或 Inpaint 遮罩功能，无需反复调试提示词，即可快速验证布光方案。
Works seamlessly with ControlNet or Inpaint masking functions, allowing for rapid verification of lighting setups without repetitive prompt tweaking.

⚙️ 使用建议 / Usage Recommendations

LoRA 强度 (Strength)：推荐 0.8 – 1.0（高权重确保色块指令被严格执行 / High weight ensures strict execution of color block instructions）。

操作方式 (Workflow)：

1. 在绘图软件中创建一个纯色图层作为“光源层”。(Create a solid color layer as the "Light Source Layer".)
2. 在需要打光的位置绘制色块（如：左侧画红色代表左侧红光，右上角画白色代表右上主光）。(Paint color blocks where lighting is needed, e.g., red on the left for red side light, white on the top-right for main light.)
3. 将该图层作为 Mask 或 ControlNet 输入，配合本 LoRA 生成图像。(Input this layer as a Mask or ControlNet input along with this LoRA to generate the image.)

适用场景 (Scenarios)：Cosplay 角色布光还原、产品摄影虚拟布光、概念设计氛围测试、人像写真创意打光。
Cosplay character lighting recreation, virtual product photography lighting, concept design atmosphere testing, creative portrait lighting.

兼容模型 (Compatibility)：适配 Flux2Klein9B 及其衍生版本。
Compatible with Flux2Klein9B and its derivatives.

📦 版本信息 / Version Info

版本 (Version)：v1（初始发布 / Initial Release）
训练目标 (Training Goal)：学习“颜色-位置-光照”三元组的映射关系，建立从二维色块到三维光影的物理渲染逻辑。
Learning the mapping relationship of the "Color-Position-Lighting" triplet to establish physical rendering logic from 2D color blocks to 3D lighting.
训练数据 (Training Data)：包含数万组“原始图+色块标注+专业布光结果”的配对数据，覆盖单色光、双色光、多色光及复杂环境光场景。
Contains tens of thousands of paired data sets of "Original Image + Color Block Annotation + Professional Lighting Result," covering single-color, dual-color, multi-color, and complex ambient light scenarios.

📌 注意事项 / Notes

本 LoRA 侧重于光影逻辑的重建，建议在使用时保持色块边缘清晰，以获得更准确的光影边界。
This LoRA focuses on reconstructing lighting logic. It is recommended to keep the edges of color blocks sharp for more accurate lighting boundaries.