Sign in

syz

@syz.bsky.social
15 followers 32 following 0 posts
PostsRepliesMedia
Reposted by syz
Marco Cuturi @marcocuturi.bsky.social · 05/11/2025
We also introduce two coupling approaches advocated this summer to improve FM training: using either very large sharp Sinkhorn couplings (arxiv.org/abs/2506.05526) or, even better, semidiscrete couplings (arxiv.org/abs/2509.25519), as proposed with Alireza Mousavi-Hosseini and @syz.bsky.social
arxiv.org
On Fitting Flow Models with Large Sinkhorn Couplings
Flow models transform data gradually from one modality (e.g. noise) onto another (e.g. images). Such models are parameterized by a time-dependent velocity field, trained to fit segments connecting pai...
011
Reposted by syz
Marco Cuturi @marcocuturi.bsky.social · 05/11/2025
We have been working with Michal Klein on pushing a module to train *flow matching* models using JAX. This is shipped as part of our new release of the OTT-JAX toolbox (github.com/ott-jax/ott) The tutorial to do so is here: ott-jax.readthedocs.io/tutorials/ne...
1147
Reposted by syz
David Picard @davidpicard.eurosky.social · 04/10/2025
Wow! Finally OT done on the entire training set to train a diffusion model!
0123
Reposted by syz
Marco Cuturi @marcocuturi.bsky.social · 03/10/2025
Our two phenomenal interns, Alireza Mousavi-Hosseini and Stephen Zhang @syz.bsky.social have been cooking some really cool work with Michal Klein and me over the summer. Relying on optimal transport couplings (to pick noise and data pairs) should, in principle, be helpful to guide flow matching 🧵
2307
Reposted by syz
arXiv cs.LG Machine Learning @cslg-bot.bsky.social · 01/10/2025
Alireza Mousavi-Hosseini, Stephen Y. Zhang, Michal Klein, Marco Cuturi: Flow Matching with Semidiscrete Couplings arxiv.org/abs/2509.25519 arxiv.org/pdf/2509.25519 arxiv.org/html/2509.25519
022