This PR improves the model card for **Wan-Alpha: High-Quality Text-to-Video Generation with Alpha Channel** by:
- Adding `tags: - rgba - transparency` to the YAML metadata, enhancing discoverability for models with alpha channel and transparency capabilities.
- Adding the official Hugging Face paper badge.
- Including the full paper abstract for better understanding of the model's contribution.
- Integrating the comprehensive "🚀 Quick Start" section, including environment setup, model download, and detailed usage instructions with a code snippet from the GitHub repository, replacing the previous simple link. This provides immediate, runnable examples for users.
- Incorporating the "🔨 Official ComfyUI Version" section, detailing how to use the model within ComfyUI.
No `library_name` tag has been added as the provided code snippets in the GitHub README do not explicitly show usage of any Hugging Face library like `diffusers`.
<div align="center">
<h1>
Wan-Alpha
</h1>
<h3>Wan-Alpha: High-Quality Text-to-Video Generation with Alpha Channel</h3>
[](https://arxiv.org/abs/)
[](https://www.xxxx)
[](https://huggingface.co/xxxx)
</div>
<img src="assets/teaser.png" alt="Wan-Alpha Qualitative Results" style="max-width: 100%; height: auto;">
>Qualitative results of video generation using **Wan-Alpha**. Our model successfully generates various scenes with accurate and clearly rendered transparency. Notably, it can synthesize diverse semi-transparent objects, glowing effects, and fine-grained details such as hair.
---
## 🔥 News
* **[2025.09.30]** Released Wan-Alpha v1.0, the Wan2.1-14B-T2V–adapted weights and inference code are now open-sourced.
---
## 🌟 Showcase
### Text-to-Video Genera…
