Sync model card with GitHub README

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+flowerteaser.png filter=lfs diff=lfs merge=lfs -text

README.md

@@ -8,62 +8,120 @@ tags:
 - bimamba
 ---
 
-# FlowerDance
+<h1 align="center">🌸 FlowerDance</h1>
+<h3 align="center">MeanFlow for Efficient and Refined 3D Dance Generation</h3>
+
+<p align="center">
+  <a href="https://arxiv.org/abs/2511.21029">
+    <img src="https://img.shields.io/badge/arXiv-FlowerDance-green" alt="Paper">
+  </a>
+  <a href="https://sun-happy-ykx.github.io/FlowerDance/">
+    <img src="https://img.shields.io/badge/Project_Page-FlowerDance-blue" alt="Project Page">
+  </a>
+  <a href="#code">
+    <img src="https://img.shields.io/badge/Conference-ECCV%202026-orange" alt="Conference">
+  </a>
+</p>
+
+<p align="center">
+  <img src="./flowerteaser.png" width="90%" alt="FlowerDance teaser">
+</p>
+
+> **Abstract**: Music-to-dance generation translates auditory signals into expressive human motion, yet existing approaches still struggle to balance refined 3D motion quality with strict inference budgets. FlowerDance is designed for both physically plausible, artistically expressive motion and efficient generation in speed and memory usage.
+>
+> FlowerDance combines MeanFlow with Physical Consistency Constraints for high-quality few-step sampling, and uses a lightweight non-autoregressive BiMamba backbone with Channel-Level Fusion for long-horizon music-to-dance synthesis. It also supports motion editing through time-decayed soft masking, enabling users to refine generated dance sequences interactively.
+
+🎉 **FlowerDance has been accepted to ECCV 2026!**
+✨ Training code release! ✨
 
-## MeanFlow for Efficient and Refined 3D Dance Generation
+---
+
+<a id="code"></a>
+
+## 🚀 Code
 
-**Accepted to ECCV 2026**
+### 🛠️ Set up the Environment
 
-[Paper](https://arxiv.org/abs/2511.21029) |
-[Project Page](https://sun-happy-ykx.github.io/FlowerDance/) |
-[GitHub](https://github.com/XulongT/FlowerDance)
+To set up the necessary environment for running this project, follow the steps below:
 
-![FlowerDance teaser](https://raw.githubusercontent.com/XulongT/FlowerDance/main/flowerteaser.png)
+1. **Create a new conda environment**
 
-FlowerDance generates physically plausible and artistically expressive 3D dance
-with efficient inference and memory usage. It combines MeanFlow with Physical
-Consistency Constraints for high-quality few-step sampling and uses a
-lightweight, non-autoregressive BiMamba backbone with Channel-Level Fusion.
+   ```bash
+   conda create -n Flower_env python=3.10
+   conda activate Flower_env
+   ```
 
-## Checkpoint
+2. **Install PyTorch (CUDA 12.8)**
 
-This repository provides the evaluation checkpoint released with FlowerDance:
+   ```
+   pip install torch==2.7.1+cu128 torchvision==0.22.1+cu128 torchaudio==2.7.1+cu128 \
+       --index-url https://download.pytorch.org/whl/cu128
+   ```
+   
+3. **Install remaining dependencies**
 
-- `train-3700.pt`
+   ```bash
+   pip install -r requirements.txt
+   ```
 
-Download it with:
+---
+
+## 📦 Download Resources
 
-```python
-from huggingface_hub import hf_hub_download
+- Download the **Preprocessed feature** from [Google Drive](https://drive.google.com/file/d/1UfCsOYMRsJAsH1LOxrJg8X5o3MRAwM2s/view?usp=sharing) and place them into `./data/` folder.
+- Download the **Checkpoints for evaluation** and place them into the `./runs/` folder:  
+  [Download Link](https://drive.google.com/file/d/1zZs_sXJToD5UzOA_m_DEoC0M79rEnEkg/view?usp=sharing)
+
+---
+
+## 🧩 Directory Structure
+
+After downloading the necessary files, ensure the directory structure follows the pattern below:
 
-checkpoint_path = hf_hub_download(
-    repo_id="xlt99/FlowerDance",
-    filename="train-3700.pt",
-)
-print(checkpoint_path)
 ```
+FlowerDance/
+    │                
+    ├── data/                 
+    ├── dataset/             
+    ├── model/                               
+    ├── runs/  
+    ├── requirements.txt
+    ├── args.py  
+    ├── EDGE.py
+    ├── inpaint.py
+    ├── test.py
+    └── vis.py     
+```
+---
 
-To use the checkpoint with the
-[official code](https://github.com/XulongT/FlowerDance), place it at:
+## 🏋️ Training
 
-```text
-runs/train/uniform2/weights/train-3700.pt
+```bash
+export WANDB_MODE=offline
+accelerate launch train.py --batch_size 128  --epochs 4000 --feature_type baseline
 ```
+---
+
+## 📏 Evaluation
+
+### 🧪 Evaluate the Model
+
+To evaluate the our model’s performance:
+
+```bash
+python test.py --batch_size 128
+```
+
 
-Then follow the evaluation instructions in the GitHub README.
+---
 
-## Citation
+## 📄 Citation
 
 ```bibtex
-@inproceedings{yang2026flowerdance,
+@article{yang2025flowerdance,
   title={FlowerDance: MeanFlow for Efficient and Refined 3D Dance Generation},
   author={Kaixing Yang and Xulong Tang and Ziqiao Peng and Xiangyue Zhang and Puwei Wang and Jun He and Hongyan Liu},
-  booktitle={European Conference on Computer Vision (ECCV)},
-  year={2026}
+  journal={arXiv preprint arXiv:2511.21029},
+  year={2025}
 }
 ```
-
-## License
-
-The release does not currently declare a license. Please contact the authors
-for permitted uses.