Wan2.1 T2V 1.3B

🚀 Wan2.1

Wan2.1 is a comprehensive and open suite of video foundation models that pushes the boundaries of video generation. It offers SOTA performance, supports consumer - grade GPUs, and excels in multiple tasks like text - to - video, image - to - video, etc.

🚀 Quick Start

📦 Installation

First, clone the repository:

git clone https://github.com/Wan-Video/Wan2.1.git
cd Wan2.1

Then, install the dependencies:

# Ensure torch >= 2.4.0
pip install -r requirements.txt

📥 Model Download

⚠️ Important Note

The 1.3B model is capable of generating videos at 720P resolution. However, due to limited training at this resolution, the results are generally less stable compared to 480P. For optimal performance, we recommend using 480P resolution.

You can download models using 🤗 huggingface - cli:

pip install "huggingface_hub[cli]"
huggingface-cli download Wan-AI/Wan2.1-T2V-1.3B --local-dir ./Wan2.1-T2V-1.3B

Or using 🤖 modelscope - cli:

pip install modelscope
modelscope download Wan-AI/Wan2.1-T2V-1.3B --local-dir ./Wan2.1-T2V-1.3B

💻 Run Text - to - Video Generation

This repository supports two Text - to - Video models (1.3B and 14B) and two resolutions (480P and 720P). The parameters and configurations for these models are as follows:

(1) Without Prompt Extention

• Single - GPU inference

python generate.py  --task t2v-1.3B --size 832*480 --ckpt_dir ./Wan2.1-T2V-1.3B --sample_shift 8 --sample_guide_scale 6 --prompt "Two anthropomorphic cats in comfy boxing gear and bright gloves fight intensely on a spotlighted stage."

If you encounter OOM (Out - of - Memory) issues, you can use the --offload_model True and --t5_cpu options to reduce GPU memory usage. For example, on an RTX 4090 GPU:

python generate.py  --task t2v-1.3B --size 832*480 --ckpt_dir ./Wan2.1-T2V-1.3B --offload_model True --t5_cpu --sample_shift 8 --sample_guide_scale 6 --prompt "Two anthropomorphic cats in comfy boxing gear and bright gloves fight intensely on a spotlighted stage."

💡 Usage Tip

If you are using the T2V - 1.3B model, we recommend setting the parameter --sample_guide_scale 6. The --sample_shift parameter can be adjusted within the range of 8 to 12 based on the performance.

• Multi - GPU inference using FSDP + xDiT USP

pip install "xfuser>=0.4.1"
torchrun --nproc_per_node=8 generate.py --task t2v-1.3B --size 832*480 --ckpt_dir ./Wan2.1-T2V-1.3B --dit_fsdp --t5_fsdp --ulysses_size 8 --sample_shift 8 --sample_guide_scale 6 --prompt "Two anthropomorphic cats in comfy boxing gear and bright gloves fight intensely on a spotlighted stage."

(2) Using Prompt Extention

Extending the prompts can effectively enrich the details in the generated videos, further enhancing the video quality. Therefore, we recommend enabling prompt extension. We provide the following two methods for prompt extension:

• Use the Dashscope API for extension.
  • Apply for a dashscope.api_key in advance (EN | CN).
  • Configure the environment variable DASH_API_KEY to specify the Dashscope API key. For users of Alibaba Cloud's international site, you also need to set the environment variable DASH_API_URL to 'https://dashscope-intl.aliyuncs.com/api/v1'. For more detailed instructions, please refer to the dashscope document.
  • Use the qwen - plus model for text - to - video tasks and qwen - vl - max for image - to - video tasks.
  • You can modify the model used for extension with the parameter --prompt_extend_model. For example:

DASH_API_KEY=your_key python generate.py  --task t2v-1.3B --size 832*480 --ckpt_dir ./Wan2.1-T2V-1.3B --prompt "Two anthropomorphic cats in comfy boxing gear and bright gloves fight intensely on a spotlighted stage" --use_prompt_extend --prompt_extend_method 'dashscope' --prompt_extend_target_lang 'ch'

• Using a local model for extension.
  • By default, the Qwen model on HuggingFace is used for this extension. Users can choose based on the available GPU memory size.
  • For text - to - video tasks, you can use models like Qwen/Qwen2.5 - 14B - Instruct, Qwen/Qwen2.5 - 7B - Instruct and Qwen/Qwen2.5 - 3B - Instruct
  • For image - to - video tasks, you can use models like Qwen/Qwen2.5 - VL - 7B - Instruct and Qwen/Qwen2.5 - VL - 3B - Instruct.
  • Larger models generally provide better extension results but require more GPU memory.
  • You can modify the model used for extension with the parameter --prompt_extend_model , allowing you to specify either a local model path or a Hugging Face model. For example:

python generate.py  --task t2v-1.3B --size 832*480 --ckpt_dir ./Wan2.1-T2V-1.3B --prompt "Two anthropomorphic cats in comfy boxing gear and bright gloves fight intensely on a spotlighted stage" --use_prompt_extend --prompt_extend_method 'local_qwen' --prompt_extend_target_lang 'ch'

(3) Runing local gradio

cd gradio
# if one uses dashscope’s API for prompt extension
DASH_API_KEY=your_key python t2v_1.3B_singleGPU.py --prompt_extend_method 'dashscope' --ckpt_dir ./Wan2.1-T2V-1.3B

# if one uses a local model for prompt extension
python t2v_1.3B_singleGPU.py --prompt_extend_method 'local_qwen' --ckpt_dir ./Wan2.1-T2V-1.3B

✨ Features

• 👍 SOTA Performance: Wan2.1 consistently outperforms existing open - source models and state - of - the - art commercial solutions across multiple benchmarks.
• 👍 Supports Consumer - grade GPUs: The T2V - 1.3B model requires only 8.19 GB VRAM, making it compatible with almost all consumer - grade GPUs. It can generate a 5 - second 480P video on an RTX 4090 in about 4 minutes (without optimization techniques like quantization). Its performance is even comparable to some closed - source models.
• 👍 Multiple Tasks: Wan2.1 excels in Text - to - Video, Image - to - Video, Video Editing, Text - to - Image, and Video - to - Audio, advancing the field of video generation.
• 👍 Visual Text Generation: Wan2.1 is the first video model capable of generating both Chinese and English text, featuring robust text generation that enhances its practical applications.
• 👍 Powerful Video VAE: Wan - VAE delivers exceptional efficiency and performance, encoding and decoding 1080P videos of any length while preserving temporal information, making it an ideal foundation for video and image generation.

📚 Documentation

Video Demos

Evaluation

We employ our Wan - Bench framework to evaluate the performance of the T2V - 1.3B model, with the results displayed in the table below. The results indicate that our smaller 1.3B model surpasses the overall metrics of larger open - source models, demonstrating the effectiveness of WanX2.1's architecture and the data construction pipeline.

Computational Efficiency on Different GPUs

We test the computational efficiency of different Wan2.1 models on different GPUs in the following table. The results are presented in the format: Total time (s) / peak GPU memory (GB).

⚠️ Important Note

(1) For the 1.3B model on 8 GPUs, set --ring_size 8 and --ulysses_size 1; (2) For the 14B model on 1 GPU, use --offload_model True; (3) For the 1.3B model on a single 4090 GPU, set --offload_model True --t5_cpu; (4) For all testings, no prompt extension was applied, meaning --use_prompt_extend was not enabled.

Introduction of Wan2.1

Wan2.1 is designed on the mainstream diffusion transformer paradigm, achieving significant advancements in generative capabilities through a series of innovations. These include our novel spatio - temporal variational autoencoder (VAE), scalable training strategies, large - scale data construction, and automated evaluation metrics. Collectively, these contributions enhance the model’s performance and versatility.

(1) 3D Variational Autoencoders

We propose a novel 3D causal VAE architecture, termed Wan - VAE specifically designed for video generation. By combining multiple strategies, we improve spatio - temporal compression, reduce memory usage, and ensure temporal causality. Wan - VAE demonstrates significant advantages in performance efficiency compared to other open - source VAEs. Furthermore, our Wan - VAE can encode and decode unlimited - length 1080P videos without losing historical temporal information, making it particularly well - suited for video generation tasks.

(2) Video Diffusion DiT

Wan2.1 is designed using the Flow Matching framework within the paradigm of mainstream Diffusion Transformers. Our model's architecture uses the T5 Encoder to encode multilingual text input, with cross - attention in each transformer block embedding the text into the model structure. Additionally, we employ an MLP with a Linear layer and a SiLU layer to process the input time embeddings and predict six modulation parameters individually. This MLP is shared across all transformer blocks, with each block learning a distinct set of biases. Our experimental findings reveal a significant performance improvement with this approach at the same parameter scale.

Data

We curated and deduplicated a candidate dataset comprising a vast amount of image and video data. During the data curation process, we designed a four - step data cleaning process, focusing on fundamental dimensions, visual quality and motion quality. Through the robust data processing pipeline, we can easily obtain high - quality, diverse, and large - scale training sets of images and videos.

Comparisons to SOTA

We compared Wan2.1 with leading open - source and closed - source models to evaluate the performace. Using our carefully designed set of 1,035 internal prompts, we tested across 14 major dimensions and 26 sub - dimensions. Then we calculated the total score through a weighted average based on the importance of each dimension. The detailed results are shown in the table below. These results demonstrate our model's superior performance compared to both open - source and closed - source models.

🔧 Technical Details

3D Variational Autoencoders

We propose a novel 3D causal VAE architecture, termed Wan - VAE specifically designed for video generation. By combining multiple strategies, we improve spatio - temporal compression, reduce memory usage, and ensure temporal causality. Wan - VAE demonstrates significant advantages in performance efficiency compared to other open - source VAEs. Furthermore, our Wan - VAE can encode and decode unlimited - length 1080P videos without losing historical temporal information, making it particularly well - suited for video generation tasks.

Video Diffusion DiT

Wan2.1 is designed using the Flow Matching framework within the paradigm of mainstream Diffusion Transformers. Our model's architecture uses the T5 Encoder to encode multilingual text input, with cross - attention in each transformer block embedding the text into the model structure. Additionally, we employ an MLP with a Linear layer and a SiLU layer to process the input time embeddings and predict six modulation parameters individually. This MLP is shared across all transformer blocks, with each block learning a distinct set of biases. Our experimental findings reveal a significant performance improvement with this approach at the same parameter scale.

📄 License

The models in this repository are licensed under the Apache 2.0 License. We claim no rights over the your generate contents, granting you the freedom to use them while ensuring that your usage complies with the provisions of this license. You are fully accountable for your use of the models, which must not involve sharing any content that violates applicable laws, causes harm to individuals or groups, disseminates personal information intended for harm, spreads misinformation, or targets vulnerable populations. For a complete list of restrictions and details regarding your rights, please refer to the full text of the license.

Acknowledgements

We would like to thank the contributors to the SD3, Qwen, umt5 - xxl, diffusers and HuggingFace repositories, for their open research.

Contact Us

If you would like to leave a message to our research or product teams, feel free to join our Discord or WeChat groups!