esfrankel.bsky.social
Eric Frankel
@esfrankel.bsky.social
83 followers 85 following 7 posts
PhD student @uwcse
Posts Media Videos Starter Packs
Pinned
esfrankel.bsky.social
Eric Frankel @esfrankel.bsky.social · Feb 27
Want to quickly sample high-quality images from diffusion models, but can’t afford the time or compute to distill them? Introducing S4S, or Solving for the Solver, which learns the coefficients and discretization steps for a DM solver to improve few-NFE generation.

Thread 👇 1/
1 1 3
esfrankel.bsky.social
Eric Frankel @esfrankel.bsky.social · Feb 27
This was a delightful collaboration with Sitan Chen, @jerryzli.bsky.social, @pangwei.bsky.social, and my wonderful advisors Lillian J. Ratliff and
@sewoong79.bsky.social. There’s lots of remaining work to be done on learning to sample that we’re excited to explore!

arxiv.org/abs/2502.17423

7/
S4S: Solving for a Diffusion Model Solver
Diffusion models (DMs) create samples from a data distribution by starting from random noise and iteratively solving a reverse-time ordinary differential equation (ODE). Because each step in the itera...
arxiv.org
1
esfrankel.bsky.social
Eric Frankel @esfrankel.bsky.social · Feb 27
S4S-Alt achieves a 1.5-2x improvement over a “vanilla” diffusion ODE solver, particularly with few NFEs, and achieves SOTA performance for techniques that do not modify the underlying DM, while being competitive with training-based methods for a fraction of a compute cost!

6/
1
esfrankel.bsky.social
Eric Frankel @esfrankel.bsky.social · Feb 27
Furthermore, to fully explore the DM solver design space, we learn both the discretization steps and the solver coefficients with S4S-Alt, which proceeds through alternating minimization, giving even further quality improvements.

5/
1
esfrankel.bsky.social
Eric Frankel @esfrankel.bsky.social · Feb 27
In S4S, we directly learn the coefficients for any diffusion solver by minimizing the global error in the final generated sample. S4S consistently outperforms traditional ODE solvers across model architectures and time discretizations, especially in the few-step regime.

4/
1
esfrankel.bsky.social
Eric Frankel @esfrankel.bsky.social · Feb 27
Other prior work fixes the ODE solver and focuses on selecting optimal discretization steps for the reverse process. While this leads to impressive gains, it overlooks the full design space of the diffusion model solver.

3/
1
esfrankel.bsky.social
Eric Frankel @esfrankel.bsky.social · Feb 27
Previous work on solving diffusion ODEs focuses on minimizing numerical errors at each step of sequentially solving the ODE. This approach works well with many steps, but breaks down when the number of steps is small.

2/
1
esfrankel.bsky.social
Eric Frankel @esfrankel.bsky.social · Feb 27
Want to quickly sample high-quality images from diffusion models, but can’t afford the time or compute to distill them? Introducing S4S, or Solving for the Solver, which learns the coefficients and discretization steps for a DM solver to improve few-NFE generation.

Thread 👇 1/
1 1 3