Text2Video Model Creation with Livepeer
Help community to build a video generation model that prefectly fits its need and enable them to use AI to generate more and more contents on Livepeer's platform.
Demo
About
The problem it solves FLock addresses the challenge of decentralized generative model development, particularly in the realm of text-to-video conversion. It leverages community collaboration for the improvement and fine-tuning of these models by enabling users to contribute prompts, generate videos, and provide feedback. By integrating Livepeer's decentralized streaming infrastructure, FLock ensures a scalable, cost-effective solution for hosting and streaming the generated videos, facilitating a transparent and democratic feedback and voting process.
How it's made
Challenges we ran into **User Engagement**: Maintaining high levels of community engagement and participation to continually improve the models. **Feedback Implementation**: Effectively incorporating community feedback into model fine-tuning in a way that genuinely improves the text-to-video models. **Decentralization**: Maintaining the decentralized nature of the platform while ensuring reliability, performance, and user experience on par with centralized solutions.