Lennart Redl

BSc Mathematics student at TUM, working on deep generative models.
Interested in both the mathematical theory and applications in the sciences.

[Lennart Redl @LennartRedl]

Going to be attending ICLR in Brazil this week to present our work on Energy-Weighted Flow Matching at the DeLTa workshop. We reformulate flow matching via importance sampling to train CNFs for Boltzmann sampling from energy evaluations.
Happy to connect! 🇧🇷

[Andreas Bergmeister @AndBergmeister]

Pretraining diffusion/flow-matching is simple and scales: noise a clean sample, regress against a closed-form target.

RL post-training can be made just as simple.

Reinforce Adjoint Matching (RAM) matches Flow-GRPO on SD3.5M in 50× fewer steps. Highest reward, no reward hacking.

[Lennart Redl @LennartRedl]

Joint work with @NiclasDern, Sebastian Pfister, @MarcelKollovieh, @ludke_david, and @guennemann. Genuinely grateful for this team 🙌
📄 https://t.co/MpGerW5ogr
📍 Mon Apr 27, 10:10–11:00, DeLTa workshop

[Lennart Redl @LennartRedl]

Going to be attending ICLR in Brazil this week to present our work on Energy-Weighted Flow Matching at the DeLTa workshop. We reformulate flow matching via importance sampling to train CNFs for Boltzmann sampling from energy evaluations.
Happy to connect! 🇧🇷