Evolution of Diffusion Video Generation Methods

Over the past couple of years, diffusion models have made remarkable progress in video generation. The recent announcement of OpenAI's Sora model has shown a whole new level of quality of generated content.

We will look at the basic principles of diffusion models for video generation and analyse how they managed to achieve such a level of quality in such a short period of time.

The talk is intended for those who have a basic understanding of the work of neural networks: convolutional and transformer.