← Browse

Europe 2026

You Might Not Need 50 Diffusion Steps — Ziv Ilan, Nvidia

Ziv Ilan

Overview

This talk explores techniques to accelerate diffusion models for image and video generation, addressing the challenge of high latency in practical applications. The core thesis is that by employing methods like quantization, caching, and distillation, developers can significantly reduce the number of diffusion steps required, thereby improving performance and enabling real-time generation without sacrificing quality.

Who should watch

Key takeaways

Notable quotes

*The main challenge once we have some interesting use cases is how to make it actually usable, right?*
*If you are able to reduce this significant number of steps, but maintaining the quality, it's something that can give you 10x, 200x improvement in performance.*
*All of it is incremental. You can use this plus this plus this. You don't necessarily need to decide, okay, I'm doing only distillation or only quantization.*

Watch on YouTube →

Unofficial community note. Prefer the recording for nuance.