Posted 3/5/2024, 10:39:07 AM
Chinese Researchers Launch Open-Source Text-to-Video Project, Aiming to Democratize AI Video Generation
- Chinese researchers from Peking University and AI company Rabbitpre starting Open-Sora project to replicate OpenAI's Sora text-to-video model
- Goal is a more user-friendly, scalable version through open-source community collaboration
- Project has made progress - developed 3-part framework and 4 video demos ranging 3-24 seconds
- Future objectives include higher resolution and longer videos, more data and computing power
- Varied reactions in China - some eager to use, others concerned about US export controls limiting access to high-performance GPUs needed