🧠 Model Overview | What It Does

MaterialShift_Klein_3D Material Transfer is a material transfer LoRA model based on Flux.2 Klein, specifically designed to solve the community issue of models only supporting surface-level texture replacement.

While preserving the original clothing's 3D structure, silhouette, and lighting logic, this model can fully transfer the material characteristics from Image 2 (weave, texture, reflectivity, thickness, softness/stiffness) onto the target category in Image 1.

The training dataset focuses primarily on apparel materials, making its performance particularly stable on 3D clothing categories like jackets, coats, and pants. At the same time, the model possesses cross-category generalization capabilities and can be applied to objects with clear material attributes, such as furniture and soft furnishings.

✨ Key Features

🎨 True 3D Material Transfer
Instead of simple texture swapping, it synchronously transfers the material's sense of thickness, wrinkle feedback, reflectivity, and shadow relationships.

💡 Apparel-Focused Training for More Stable Results
Extensive training data on apparel materials makes the model output more natural regarding clothing structure and body force logic.

📸 Preserves Original Structure & Composition
Maintains the silhouette, seams, proportions, and pose of Image 1, altering only the material rendering.

🖼 Cross-Category Transfer Support
Beyond apparel, it can be extended to objects with distinct material attributes, such as furniture and soft furnishings.

🔑 Trigger Prompt

Change the material of the [target category] in Image 1 to match the material in Image 2.

It is recommended to specify [target category] to a concrete object, such as: jacket / coat / sofa / chair / pants

Reprinted from: MaterialShift_Klein_Material Transfer