We welcome you to CDS-KIAC talk on 17th September 2025 (Wednesday). The details are as below:
Speaker: Sayak Paul, Hugging Face
Title: State of Open Video Generation Models
Date and Time: September 17, 2025: 3:00 PM
Venue: #102, CDS Seminar Hall.
Abstract: In this session, we will cover the state of open video generation models. For the past few years, the GenAI community has seen an emergence in photorealistic image generation models like Flux, Nano Banana, and so on. 2025 is gradually setting itself up for videos. With a fair amount of divide between open and closed models for video generation, it can be daunting to even think about the possibility of open high-quality video generation models. This session will try to give wings to those possibilities by showing what the open video generation community has been up to. We will discuss trends in the architectures, tiny and neat training techniques. This talk will include both inference and fine-tuning.
Bio of Speaker: Sayak works on diffusion models at Hugging Face. His day-to-day includes contributing to the diffusers library, training and babysitting diffusion models. He’s interested in subject-driven generation, preference alignment, and inference-time scaling of diffusion models. When he is not working, he can be found playing the guitar and binge-watching ICML tutorials and Suits.
Host Faculty: Dr. Anirban Chakraborty
ALL ARE WELCOME
