v1.0.0

joshiayush/inbv1.0.0Mar 17, 2024by FrankLeeeee

AI Summary

The initial version of Open-Sora is released, providing open-source source code for Sora-like video generation, optimized training infrastructure via ColossalAI, and pre-trained model weights.

Key Highlights

Open-source Sora-like video generation pipeline
System optimization using ColossalAI and Flash Attention
Pre-trained model weights for specific resolutions
Source code for data processing

New Features

Sora-like video generation training and inference pipeline
ColossalAI and Flash Attention optimization
Pre-trained model weights (16x256x256, 16x512x512)
Data processing source code

Full Release Notes

We present Open-Sora, an initiative dedicated to efficiently produce high-quality video and make the model, tools and contents accessible to all. By embracing open-source principles, Open-Sora not only democratizes access to advanced video generation techniques, but also offers a streamlined and user-friendly platform that simplifies the complexities of video production. With Open-Sora, we aim to inspire innovation, creativity, and inclusivity in the realm of content creation.

In this release, we have included
- source code for data processing
- source code for sora-like video generation training and inference pipeline
- system optimization powered by ColossalAI and flash attention
- model weights trained for 16×256×256 and 16×512x512 resolutions, available on Hugging Face