ShiguiLi / EVODiffLinks

[NeurIPS 2025🔥:] EVODiff is an inference-time refinement method for diffusion models that improves sampling efficiency and generative fidelity by systematically reducing conditional entropy, without relying on reference trajectories.
28Updated last week

Alternatives and similar repositories for EVODiff

Users that are interested in EVODiff are comparing it to the libraries listed below

Sorting: