Update README.md

README.md

@@ -1,9 +1,14 @@
 ---
 license: cc-by-nc-sa-4.0
 ---
-# MagicTryOn: Harnessing Diffusion Transformer for Garment-Preserving Video Virtual Try-on
+<h2 align="center"> 
+  <a href="https://arxiv.org/abs/2505.21325v2">
+    MagicTryOn: Harnessing Diffusion Transformer for Garment-Preserving Video Virtual Try-on
+  </a>
+</h2>
 
 <a href="https://arxiv.org/abs/2505.21325v2"><img src='https://img.shields.io/badge/arXiv-2501.11325-red?style=flat&logo=arXiv&logoColor=red' alt='arxiv'></a>&nbsp;
+<a href="https://huggingface.co/LuckyLiGY/MagicTryOn"><img src='https://img.shields.io/badge/Hugging Face-ckpts-orange?style=flat&logo=HuggingFace&logoColor=orange' alt='huggingface'></a>&nbsp;
 <a href="https://vivocameraresearch.github.io/magictryon/"><img src='https://img.shields.io/badge/Project-Page-Green' alt='GitHub'></a>&nbsp;
 <a href="http://www.apache.org/licenses/LICENSE-2.0"><img src='https://img.shields.io/badge/License-CC BY--NC--SA--4.0-lightgreen?style=flat&logo=Lisence' alt='License'></a>&nbsp;
 
@@ -13,11 +18,11 @@ license: cc-by-nc-sa-4.0
   <img src="asset/model.png" width="100%" height="100%"/>
 </div>
 
-## Updates 
-- **`2025/06/06`**: 🎉 We are excited to announce that the ***code and weights*** of [**MagicTryOn**](https://github.com/vivoCameraResearch/Magic-TryOn/) have been released! Check it out! You can download the weights from 🤗[**HuggingFace**](https://huggingface.co/LuckyLiGY/MagicTryOn).
+## 📣 News 
+- **`2025/06/09`**: 🎉 We are excited to announce that the ***code*** of [**MagicTryOn**](https://github.com/vivoCameraResearch/Magic-TryOn/) have been released! Check it out! ***The weights are on the way and are expected to be released on June 14***. You can download the weights from 🤗[**HuggingFace**](https://huggingface.co/LuckyLiGY/MagicTryOn) once they are open-sourced.
 - **`2025/05/27`**: Our [**Paper on ArXiv**](https://arxiv.org/abs/2505.21325v2) is available 🥳!
 
-## To-Do List for MagicTryOn Release
+## ✅ To-Do List for MagicTryOn Release
 - ✅ Release the source code
 - ✅ Release the inference demo and pretrained weights
 - ✅ Release the customized try-on utilities
@@ -26,7 +31,7 @@ license: cc-by-nc-sa-4.0
 - [  ] Release the second version of the pretrained model weights 
 - [  ] Update Gradio App. 
 
-## Installation
+## 😍 Installation
 
 Create a conda environment & Install requirments 
 ```shell
@@ -37,12 +42,17 @@ pip install -r requirements.txt
 # or
 conda env create -f environment.yaml
 ```
-If you encounter an error while installing Flash Attention, please [**manually download**](https://github.com/Dao-AILab/flash-attention/releases) the installation package based on your Python version, CUDA version, and Torch version, and install it using ***pip install***.
+If you encounter an error while installing Flash Attention, please [**manually download**](https://github.com/Dao-AILab/flash-attention/releases) the installation package based on your Python version, CUDA version, and Torch version, and install it using `pip install flash_attn-2.7.3+cu12torch2.2cxx11abiFALSE-cp312-cp312-linux_x86_64.whl`.
 
+Use the following command to download the weights:
+```PowerShell
+cd Magic-TryOn
+HF_ENDPOINT=https://hf-mirror.com huggingface-cli download LuckyLiGY/MagicTryOn --local-dir ./weights/MagicTryOn_14B_V1
+```
 
-## Demo Inference
+## 😉 Demo Inference
 ### 1. Image TryOn
-You can directly run the following command to perform image try-on. If you want to modify some inference parameters, please make the changes inside the ***predict_image_tryon_up.py*** file.
+You can directly run the following command to perform image try-on demo. If you want to modify some inference parameters, please make the changes inside the `predict_image_tryon_up.py` file.
 ```PowerShell
 CUDA_VISIBLE_DEVICES=0 python predict_image_tryon_up.py
 
@@ -50,7 +60,7 @@ CUDA_VISIBLE_DEVICES=1 python predict_image_tryon_low.py
 ```
 
 ### 2. Video TryOn
-You can directly run the following command to perform image try-on. If you want to modify some inference parameters, please make the changes inside the ***predict_video_tryon_up.py*** file.
+You can directly run the following command to perform image try-on demo. If you want to modify some inference parameters, please make the changes inside the `predict_video_tryon_up.py` file.
 ```PowerShell
 CUDA_VISIBLE_DEVICES=0 python predict_video_tryon_up.py
 
@@ -67,17 +77,16 @@ Before performing customized try-on, you need to complete the following five ste
     ```
 
 2. **Cloth Line Map**  
-   Extract the structural lines or sketch of the garment using [**AniLines-Anime-Lineart-Extractor**](https://github.com/zhenglinpan/AniLines-Anime-Lineart-Extractor).
-
+   Extract the structural lines or sketch of the garment using [**AniLines-Anime-Lineart-Extractor**](https://github.com/zhenglinpan/AniLines-Anime-Lineart-Extractor). Download the pre-trained models from this [**link**](https://drive.google.com/file/d/1oazs4_X1Hppj-k9uqPD0HXWHEQLb9tNR/view?usp=sharing) and put them in the `inference/customize/AniLines/weights` folder.
    ```PowerShell
     cd inference/customize/AniLines
     python infer.py --dir_in datasets/garment/vivo/vivo_garment --dir_out datasets/garment/vivo/vivo_garment_anilines --mode detail --binarize -1 --fp16 True --device cuda:1
     ```
 
 3. **Mask**  
-   Generate the agnostic mask of the garment, which is essential for region control during try-on. Please [**download**]() the required checkpoint for obtaining the agnostic mask. The checkpoint needs to be placed in the ***inference/customize/gen_mask/ckpt*** folder.
+   Generate the agnostic mask of the garment, which is essential for region control during try-on. Please [**download**](https://drive.google.com/file/d/1E2JC_650g69AYrN2ZCwc8oz8qYRo5t5s/view?usp=sharing) the required checkpoint for obtaining the agnostic mask. The checkpoint needs to be placed in the `inference/customize/gen_mask/ckpt` folder.
 
-   (1) You need to rename your video to ***video.mp4***, and then construct the folders according to the following directory structure.
+   (1) You need to rename your video to `video.mp4`, and then construct the folders according to the following directory structure.
     ```
     ├── datasets
     │   ├── person
@@ -93,7 +102,7 @@ Before performing customized try-on, you need to complete the following five ste
     |   |   |   |   ├── 00002 ...
     ```
 
-    (2) Using ***video2image.py*** to convert the video into image frames and save them to ***00001/images***.
+    (2) Using `video2image.py` to convert the video into image frames and save them to `datasets/person/customize/video/00001/images`.
 
     (3) Run the following command to obtain the agnostic mask.
 
@@ -107,9 +116,9 @@ Before performing customized try-on, you need to complete the following five ste
     # mask, _ = get_mask_location('dc', "dresses", model_parse, keypoints)
     ```
 
-    After completing the above steps, you will obtain the agnostic masks for all video frames in the ***00001/masks*** folder.
+    After completing the above steps, you will obtain the agnostic masks for all video frames in the `datasets/person/customize/video/00001/masks` folder.
 4. **Agnostic Representation**  
-   Construct an agnostic representation of the person by removing garment-specific features. You can directly run ***get_masked_person.py*** to obtain the Agnostic Representation. Make sure to modify the ***image_folder*** and ***mask_folder*** parameters. The resulting video frames will be stored in ***00001/agnostic***.
+   Construct an agnostic representation of the person by removing garment-specific features. You can directly run `get_masked_person.py` to obtain the Agnostic Representation. Make sure to modify the `--image_folder` and `--mask_folder` parameters. The resulting video frames will be stored in `datasets/person/customize/video/00001/agnostic`.
 
 5. **DensePose**  
    Use DensePose to obtain UV-mapped dense human body coordinates for better spatial alignment.
@@ -121,20 +130,20 @@ Before performing customized try-on, you need to complete the following five ste
     cd inference/customize/detectron2/projects/DensePose
     bash run.sh
     ```
-    (3) The generated results will be stored in the ***00001/image-densepose*** folder.
+    (3) The generated results will be stored in the `datasets/person/customize/video/00001/image-densepose` folder.
 
-After completing the above steps, run the ***image2video.py*** file to generate the required customized videos: ***mask.mp4***, ***agnostic.mp4***, and ***densepose.mp4***. Then, run the following command:
+After completing the above steps, run the `image2video.py` file to generate the required customized videos: `mask.mp4`, `agnostic.mp4`, and `densepose.mp4`. Then, run the following command:
 ```PowerShell
 CUDA_VISIBLE_DEVICES=0 python predict_video_tryon_customize.py
 ```
 
-## Acknowledgement
-Our code is modified based on [VideoX-Fun](https://github.com/aigc-apps/VideoX-Fun/tree/main). We adopt [Wan2.1-I2V-14B](https://github.com/Wan-Video/Wan2.1) as the base model. We use [SCHP](https://github.com/GoGoDuck912/Self-Correction-Human-Parsing/tree/master) and [openpose](https://github.com/CMU-Perceptual-Computing-Lab/openpose) to generate masks. We use [detectron2](https://github.com/facebookresearch/detectron2) to generate densepose. Thanks to all the contributors!
+## 😘 Acknowledgement
+Our code is modified based on [VideoX-Fun](https://github.com/aigc-apps/VideoX-Fun/tree/main). We adopt [Wan2.1-I2V-14B](https://github.com/Wan-Video/Wan2.1) as the base model. We use [SCHP](https://github.com/GoGoDuck912/Self-Correction-Human-Parsing/tree/master), [openpose](https://github.com/CMU-Perceptual-Computing-Lab/openpose), and [DensePose](https://github.com/facebookresearch/DensePose) to generate masks. We use [detectron2](https://github.com/facebookresearch/detectron2) to generate densepose. Thanks to all the contributors!
 
-## License
+## 😊 License
 All the materials, including code, checkpoints, and demo, are made available under the [Creative Commons BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/) license. You are free to copy, redistribute, remix, transform, and build upon the project for non-commercial purposes, as long as you give appropriate credit and distribute your contributions under the same license.
 
-## Citation
+## 🤩 Citation
 
 ```bibtex
 @misc{li2025magictryon,
@@ -146,4 +155,4 @@ All the materials, including code, checkpoints, and demo, are made available und
       primaryClass={cs.CV},
       url={https://arxiv.org/abs/2505.21325}, 
 }
-```
+```