Improve model card: add paper link, authors, and teaser image

README.md

@@ -1,6 +1,8 @@
 ---
-license: mit
+base_model:
+- black-forest-labs/FLUX.1-dev
 library_name: diffusers
+license: mit
 pipeline_tag: image-to-image
 tags:
 - flux
@@ -11,17 +13,25 @@ tags:
 - depth
 - normal
 - lighting
-base_model:
-- black-forest-labs/FLUX.1-dev
 ---
 
 # DealMaTe: Multi-Dimensional Material Transfer via Diffusion Transformer
 
-These are the pre-trained **Shader LoRA weights** for [DealMaTe](https://github.com/haha-lisa/DealMaTe) — an efficient, text-free material transfer framework built on FLUX.1 Diffusion Transformer.
+These are the pre-trained **Shader LoRA weights** for [DealMaTe](https://github.com/haha-lisa/DealMaTe) — an efficient, text-free material transfer framework built on the FLUX.1 Diffusion Transformer.
+
+[**Paper**](https://huggingface.co/papers/2605.15681) | [**ArXiv**](https://arxiv.org/abs/2605.15681) | [**Code**](https://github.com/haha-lisa/DealMaTe)
+
+**Authors:** Nisha Huang, Yizhou Lin, Jie Guo, Xiu Li, Tong-Yee Lee, Zitong Yu.
+
+<p align="center">
+  <img src="https://raw.githubusercontent.com/haha-lisa/DealMaTe/main/assets/teaser.jpg" alt="DealMaTe teaser" width="900"/>
+</p>
 
 Given a material reference image and a target object image, DealMaTe faithfully transfers the material appearance onto the object while preserving its 3D geometry, lighting, and surface structure.
 
-**Paper:** *ACM Transactions on Graphics (TOG), 2026*
+## Overview
+
+DealMaTe is a simplified diffusion framework that eliminates text dependency and reference networks by using depth, normal, and lighting images for material transfer. It utilizes a lightweight 3D information injection method, **Multi-Dim 3D Shader LoRA**, which enables compatible control conditions and achieves harmonious results without modifying base model weights.
 
 ## Model Weights
 
@@ -57,7 +67,8 @@ python inference.py \
     --normal_path    examples/inputs/normal.png \
     --lighting_path  examples/inputs/lighting.png \
     --output_path    outputs/result.png \
-    --lora_path      ./lora
+    --lora_path      ./lora \
+    --seed           42
 ```
 
 **Required base models (auto-downloaded):**