Wan-Alpha_ComfyUI

Wan-Alpha: High-Quality Text-to-Video Generation with Alpha Channel

Paper Link

Abstract

RGBA video generation, which includes an alpha channel to represent transparency, is gaining increasing attention across a wide range of applications. However, existing methods often neglect visual quality, limiting their practical usability. In this paper, we propose Wan-Alpha, a new framework that generates transparent videos by learning both RGB and alpha channels jointly. We design an effective variational autoencoder (VAE) that encodes the alpha channel into the RGB latent space. Then, to support the training of our diffusion transformer, we construct a high-quality and diverse RGBA video dataset. Compared with state-to-art methods, our model demonstrates superior performance in visual quality, motion realism, and transparency rendering. Notably, our model can generate a wide variety of semi-transparent objects, glowing effects, and fine-grained details such as hair strands. The released model is available on our website: this https URL .

<div align="center"> <h1> Wan-Alpha </h1> <h3>Wan-Alpha: High-Quality Text-to-Video Generation with Alpha Channel</h3>

</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 Generation with Alpha Channel

For more results, please visit Our Website

🚀 Quick Start

1. Environment Setup

# Clone the project repository
git clone https://github.com/WeChatCV/Wan-Alpha.git
cd Wan-Alpha

# Create and activate Conda environment
conda create -n Wan-Alpha python=3.11 -y
conda activate Wan-Alpha

# Install dependencies
pip install -r requirements.txt

2. Model Download

Download Wan2.1-T2V-14B

Download Lightx2v-T2V-14B

Download Wan-Alpha VAE

🧪 Usage

You can test our model through:

torchrun --nproc_per_node=8 --master_port=29501 generate_dora_lightx2v.py --size 832*480\
         --ckpt_dir "path/to/your/Wan-2.1/Wan2.1-T2V-14B" \
         --dit_fsdp --t5_fsdp --ulysses_size 8 \
         --vae_lora_checkpoint "path/to/your/decoder.bin" \
         --lora_path "path/to/your/epoch-13-1500.safetensors" \
         --lightx2v_path "path/to/your/lightx2v_T2V_14B_cfg_step_distill_v2_lora_rank64_bf16.safetensors" \
         --sample_guide_scale 1.0 \
         --frame_num 81 \
         --sample_steps 4 \
         --lora_ratio 1.0 \
         --lora_prefix "" \
         --prompt_file ./data/prompt.txt \
         --output_dir ./output

You can specify the weights of Wan2.1-T2V-14B with --ckpt_dir, LightX2V-T2V-14B with --lightx2v_path, Wan-Alpha-VAE with --vae_lora_checkpoint, and Wan-Alpha-T2V with --lora_path. Finally, you can find the rendered RGBA videos with a checkerboard background and PNG frames at --output_dir.

Prompt Writing Tip: You need to specify that the background of the video is transparent, the visual style, the shot type (such as close-up, medium shot, wide shot, or extreme close-up), and a description of the main subject. Prompts support both Chinese and English input.

# An example of prompt.
This video has a transparent background. Close-up shot. A colorful parrot flying. Realistic style.

🔨 Official ComfyUI Version

Note: We have reorganized our models to ensure they can be easily loaded into ComfyUI. Please note that these models differ from the ones mentioned above.

1. Download models

• The Wan DiT base model: wan2.1_t2v_14B_fp16.safetensors
• The Wan text encoder: umt5_xxl_fp8_e4m3fn_scaled.safetensors
• The LightX2V model: lightx2v_T2V_14B_cfg_step_distill_v2_lora_rank64_bf16.safetensors
• Our RGBA Dora: epoch-13-1500_changed.safetensors
• Our RGB VAE Decoder: wan_alpha_2.1_vae_rgb_channel.safetensors.safetensors
• Our Alpha VAE Decoder: wan_alpha_2.1_vae_alpha_channel.safetensors.safetensors

2. Copy the files into the ComfyUI/models folder and organize them as follows:

ComfyUI/models
├── diffusion_models
│   └── wan2.1_t2v_14B_fp16.safetensors
├── loras
│   ├── epoch-13-1500_changed.safetensors
│   └── lightx2v_T2V_14B_cfg_step_distill_v2_lora_rank64_bf16.safetensors
├── text_encoders
│   └── umt5_xxl_fp8_e4m3fn_scaled.safetensors
├── vae
│   ├── wan_alpha_2.1_vae_alpha_channel.safetensors.safetensors
│   └── wan_alpha_2.1_vae_rgb_channel.safetensors.safetensors

3. Install our custom RGBA video previewer and PNG frames zip packer. Copy the file RGBA_save_tools.py into the ComfyUI/custom_nodes folder.

• Thanks to @mr-lab for an improved WebP version! You can find it in this issue.

4. Example workflow: wan_alpha_t2v_14B.json

<img src="comfyui/comfyui.jpg" style="margin:auto;"/>

---

🤝 Acknowledgements

This project is built upon the following excellent open-source projects:

• DiffSynth-Studio (training/inference framework)
• Wan2.1 (base video generation model)
• LightX2V (inference acceleration)
• WanVideo_comfy (inference acceleration)

We sincerely thank the authors and contributors of these projects.

---

✏ Citation

If you find our work helpful for your research, please consider citing our paper:

@misc{dong2025wanalpha,
      title={Wan-Alpha: High-Quality Text-to-Video Generation with Alpha Channel}, 
      author={Haotian Dong and Wenjing Wang and Chen Li and Di Lin},
      year={2025},
      eprint={2509.24979},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2509.24979}, 
}

---

📬 Contact Us

If you have any questions or suggestions, feel free to reach out via GitHub Issues . We look forward to your feedback!