Adverserial-Flow-Matching-with-Learned-Prior. Extension of adversarial flow matching where the noise distribution (prior) is learned rather than fixed to N(0, I). This allows the model to potentially find a better matching between the prior and data manifold.

github.com/Algomancer/Adverserial-Flow-Matching-with-Learned-Prior

Problem

Audience

Updates

No founder updates yet. Check back soon.