Wan 2.1
Wan2.1 is an open-source suite of large-scale video generation models that supports text-to-video, image-to-video, video editing, text-to-image, and video-to-audio tasks. It aims to provide state-of-the-art performance while being accessible on consumer-grade GPUs, with the smallest model requiring only 8.19 GB VRAM. It is designed for developers and researchers who need a versatile and high-quality video generation tool.
✨ Key features
- SOTA performance across multiple benchmarks
- Consumer-grade GPU support (T2V-1.3B uses 8.19 GB VRAM)
- Multiple tasks: T2V, I2V, video editing, T2I, V2A
- First video model generating Chinese and English text
- Powerful Wan-VAE for efficient encoding/decoding of 1080P videos
- Integrations with Diffusers, ComfyUI, and Gradio
🎯 Use cases
- Generate videos from text prompts for creative projects
- Create videos from images for animation or storytelling
- Edit videos with text or image guidance
- Generate images from text descriptions
- Add audio to videos
📦 Installation
🧰 Requirements: Python environment with PyTorch >= 2.4.0; models downloaded from Hugging Face or ModelScope; GPU recommended (consumer-grade for 1.3B model).
git clone https://github.com/Wan-Video/Wan2.1.git
cd Wan2.1
Install dependencies:
# Ensure torch >= 2.4.0
pip install -r requirements.txt
🚀 Usage
python generate.py --task t2v-14B --size 1280*720 --ckpt_dir ./Wan2.1-T2V-14B --prompt "Two anthropomorphic cats in comfy boxing gear and bright gloves fight intensely on a spotlighted stage."
⚠️ Good to know
The 1.3B model is less stable at 720P resolution; first-last-frame model performs better with Chinese prompts.
❓ FAQ
What are the hardware requirements for running Wan2.1?
The T2V-1.3B model requires only 8.19 GB VRAM, making it compatible with most consumer GPUs. For larger models, multi-GPU inference is supported.
How can I download the models?
Models can be downloaded using huggingface-cli or modelscope-cli. For example, use huggingface-cli download Wan-AI/Wan2.1-T2V-14B --local-dir ./Wan2.1-T2V-14B.
Does Wan2.1 support video editing?
Yes, Wan2.1 supports video editing as one of its multiple tasks, along with text-to-video, image-to-video, text-to-image, and video-to-audio.
Can I generate videos with Chinese text?
Yes, Wan2.1 is the first video model capable of generating both Chinese and English text.
📊 Repository
🤖 Overview, features, install steps and FAQ were generated from the project's README on Sep 4, 2026. Always check the original source before running commands.