Sayak Paul

Hi there 👋 I am Sayak Paul (সায়ক পাল). I work on 🧨 diffusion models at Hugging Face. Know more about me from here. I maintain a Google Doc answering some FAQs at length. You can check it out here.

My external articles and other publishing engagements (like books, liveProjects, etc.) are listed here. Decks from my speaking engagements are listed here. A detailed account of the things that are not directly available from the top navbar (interviews, talks etc.) can be found here.

The structure of this website is inspired by Omar’s site.

News

• New work in collaboration with KAIST AI and Korea University: Margin-aware Preference Optimization for Aligning Diffusion Models without Reference.
• We open-sourced a repository diffusers-torchao that contains recipes to optimize the runtime for large diffusion models like Flux and Cog-Video.
• We open-sourced cogvideox-factory, a repository that allows fine-tuning the Cog family of T2V and I2V models under 24GB GPU VRAM.
• Shipped quantization in Diffusers. Follow this thread to know more.
• Flux.1 Dev LoRA training with quantization is now possible. Guide is here.

To know about my projects, please refer to my GitHub profile.

Apart from the blogs here, I try to contribute to other platforms in the form of writing. Please refer here for more details.