Researcher, Apple
1 paper at NeurIPS 2025
We propose a highly efficient distillation algorithm for flow matching models that requires minimal computation and enables few-shot training.
