metadata

license: mit
pipeline_tag: text-to-video

Open-Sora Plan

This project aims to create a simple and scalable repo, to reproduce Sora (OpenAI, but we prefer to call it "ClosedAI" ). We wish the open-source community can contribute to this project. Pull requests are welcome! The current code supports complete training and inference using the Huawei Ascend AI computing system. Models trained on Huawei Ascend can also output video quality comparable to industry standards.

本项目希望通过开源社区的力量复现Sora，由北大-兔展AIGC联合实验室共同发起，当前版本离目标差距仍然较大，仍需持续完善和快速迭代，欢迎Pull request！目前代码同时支持使用国产AI计算系统（华为昇腾）进行完整的训练和推理。基于昇腾训练出的模型，也可输出持平业界的视频质量。

📣 News

COMING SOON ⚡️⚡️⚡️ For large model parallelisation training, TP & SP and more strategies are coming...

近期将新增华为昇腾多模态MindSpeed-MM分支，借助华为MindSpeed-MM套件的能力支撑Open-Sora Plan参数的扩增，为更大参数规模的模型训练提供TP、SP等分布式训练能力。
[2024.10.16] 🎉 We released version 1.3.0, featuring: WFVAE, pompt refiner, data filtering strategy, sparse attention, and bucket training strategy. We also support 93x480p within 24G VRAM. More details can be found at our latest report.
[2024.08.13] 🎉 We are launching Open-Sora Plan v1.2.0 I2V model, which based on Open-Sora Plan v1.2.0. The current version supports image-to-video generation and transition generation (the starting and ending frames conditions for video generation). Checking out the Image-to-Video section in this report.
[2024.07.24] 🔥🔥🔥 v1.2.0 is here! Utilizing a 3D full attention architecture instead of 2+1D. We released a true 3D video diffusion model trained on 4s 720p. Checking out our latest report.
[2024.05.27] 🎉 We are launching Open-Sora Plan v1.1.0, which significantly improves video quality and length, and is fully open source! Please check out our latest report. Thanks to ShareGPT4Video's capability to annotate long videos.
[2024.04.09] 🤝 Excited to share our latest exploration on metamorphic time-lapse video generation: MagicTime, which learns real-world physics knowledge from time-lapse videos.
[2024.04.07] 🎉🎉🎉 Today, we are thrilled to present Open-Sora-Plan v1.0.0, which significantly enhances video generation quality and text control capabilities. See our report. Thanks to HUAWEI NPU for supporting us.
[2024.03.27] 🚀🚀🚀 We release the report of VideoCausalVAE, which supports both images and videos. We present our reconstructed video in this demonstration as follows. The text-to-video model is on the way.
[2024.03.01] 🤗 We launched a plan to reproduce Sora, called Open-Sora Plan! Welcome to watch 👀 this repository for the latest updates.

😍 Gallery

Text & Image to Video Generation.

😮 Highlights

Open-Sora Plan shows excellent performance in video generation.

🔥 High performance CausalVideoVAE, but with fewer training cost

High compression ratio with excellent performance, capable of compressing videos by 256 times (4×8×8). Causal convolution supports simultaneous inference of images and videos but only need 1 node to train.

🚀 Video Diffusion Model based on 3D attention, joint learning of spatiotemporal features.

With a new sparse attention architecture instead of a 2+1D model, 3D attention can better capture joint spatial and temporal features.

🤗 Demo

Gradio Web UI

Highly recommend trying out our web demo by the following command.

python -m opensora.serve.gradio_web_server --model_path "path/to/model" \
    --ae WFVAEModel_D8_4x8x8 --ae_path "path/to/vae" \
    --caption_refiner "path/to/refiner" \
    --text_encoder_name_1 "path/to/text_enc" --rescale_betas_zero_snr

ComfyUI

Coming soon...

🐳 Resource

Version	Architecture	Diffusion Model	CausalVideoVAE	Data	Prompt Refiner
v1.3.0	3D	Anysize in 93x640x640[3], more checkpoints are coming soon	Anysize	prompt_refiner	checkpoint
v1.2.0	3D	93x720p, 29x720p[1], 93x480p[1,2], 29x480p, 1x480p, 93x480p_i2v	Anysize	Annotations	-
v1.1.0	2+1D	221x512x512, 65x512x512	Anysize	Data and Annotations	-
v1.0.0	2+1D	65x512x512, 65x256x256, 17x256x256	Anysize	Data and Annotations	-

[1] Please note that the weights for v1.2.0 29×720p and 93×480p were trained on Panda70M and have not undergone final high-quality data fine-tuning, so they may produce watermarks.

[2] We fine-tuned 3.5k steps from 93×720p to get 93×480p for community research use.

[3] The model is trained arbitrarily on stride=32. So keep the resolution of the inference a multiple of 32. Frames needs to be 4n+1, e.g. 93, 77, 61, 45, 29, 1 (image).

🚨 For version 1.2.0, we no longer support 2+1D models.

⚙️ Requirements and Installation

Clone this repository and navigate to Open-Sora-Plan folder

git clone https://github.com/PKU-YuanGroup/Open-Sora-Plan
cd Open-Sora-Plan

Install required packages We recommend the requirements as follows.

Python >= 3.8
Pytorch >= 2.1.0
CUDA Version >= 11.7

conda create -n opensora python=3.8 -y
conda activate opensora
pip install -e .

Install optional requirements such as static type checking:

pip install -e '.[dev]'

🗝️ Training & Inferencing

🗜️ CausalVideoVAE

The data preparation, training, inferencing and evaluation can be found here

📖 Prompt Refiner

The data preparation, training, inferencing can be found here

📜 Text-to-Video

The data preparation, training and inferencing can be found here

🖼️ Image-to-Video

The data preparation, training and inferencing can be found here

⚡️ Extra Save Memory

🔆 Training

During training, the entire EMA model remains in VRAM. You can enable --offload_ema or disable --use_ema. Additionally, VAE tiling is disabled by default, but you can pass --enable_tiling or disable --vae_fp32. Finally, a temporary but extreme saving memory option is enable --extra_save_mem to offload the text encoder and VAE to the CPU when not in use, though this will significantly slow down performance.

We currently have two plans: one is to continue using the Deepspeed/FSDP approach, sharding the EMA and text encoder across ranks with Zero3, which is sufficient for training 10-15B models. The other is to adopt MindSpeed for various parallel strategies, enabling us to scale the model up to 30B.

⚡️ 24G VRAM Inferencing

Please first ensure that you understand how to inference. Refer to the inference instructions in Text-to-Video. Simply specify --save_memory, and during inference, enable_model_cpu_offload(), enable_sequential_cpu_offload(), and vae.vae.enable_tiling() will be automatically activated.

💡 How to Contribute

We greatly appreciate your contributions to the Open-Sora Plan open-source community and helping us make it even better than it is now!

For more details, please refer to the Contribution Guidelines

👍 Acknowledgement

Latte: It is an wonderful 2+1D video generated model.
PixArt-alpha: Fast Training of Diffusion Transformer for Photorealistic Text-to-Image Synthesis.
ShareGPT4Video: Improving Video Understanding and Generation with Better Captions.
VideoGPT: Video Generation using VQ-VAE and Transformers.
DiT: Scalable Diffusion Models with Transformers.
FiT: Flexible Vision Transformer for Diffusion Model.
Positional Interpolation: Extending Context Window of Large Language Models via Positional Interpolation.

🔒 License

See LICENSE for details.

✨ Star History

✏️ Citing

BibTeX

@misc{lin2024opensoraplanopensourcelarge,
      title={Open-Sora Plan: Open-Source Large Video Generation Model}, 
      author={Bin Lin and Yunyang Ge and Xinhua Cheng and Zongjian Li and Bin Zhu and Shaodong Wang and Xianyi He and Yang Ye and Shenghai Yuan and Liuhan Chen and Tanghui Jia and Junwu Zhang and Zhenyu Tang and Yatian Pang and Bin She and Cen Yan and Zhiheng Hu and Xiaoyi Dong and Lin Chen and Zhang Pan and Xing Zhou and Shaoling Dong and Yonghong Tian and Li Yuan},
      year={2024},
      eprint={2412.00131},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2412.00131}, 
}

LanguageBind
/

Open-Sora-Plan-v1.3.0