v1.0.0
datalab-to/suryav1.0.0Mar 17, 2024by FrankLeeeee
AI Summary
Open-Sora V1.0.0 introduces a comprehensive open-source platform for high-quality video generation. It provides source code for data processing and a Sora-like training and inference pipeline, optimized with ColossalAI and flash attention, alongside pre-trained model weights.
Key Highlights
- Open-source video generation initiative
- Source code for data processing and Sora-like pipelines
- Optimization powered by ColossalAI and Flash Attention
- Pre-trained model weights for 16x256x256 and 16x512x512 resolutions
New Features
- Data processing source code
- Video generation training pipeline
- Video generation inference pipeline
- ColossalAI system optimization
- Flash attention integration
- Pre-trained video model weights
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