Two-person research lab training generative text-to-video models from scratch and publishing field notes on the process.
Linum is a research lab building generative text-to-video models, training them from scratch and sharing research, releases, and real-world lessons learned in the process. The team publishes field notes documenting their development process and insights from training video generation models. It is positioned as a focused small team exploring text-to-video generation technology. Linum is a product of Linum.