<ul><li>DiverseFlow is a training-free approach to improve the diversity of flow models.</li><li>It uses a determinantal point process to induce a coupling between samples, driving diversity within a fixed sampling budget.</li><li>DiverseFlow allows exploration of more variations in a learned flow model with fewer samples, making it sample-efficient.</li><li>The method demonstrates efficacy in tasks such as text-guided image generation, large-hole inpainting, and class-conditional image synthesis.</li></ul>

DiverseFlow: Sample-Efficient Diverse Mode Coverage in Flows

Discover more