The prediction of stable crystal structures is an important part of designing solid-state crystalline materials with desired properties. Recent advances in structural feature representations and generative neural networks promise the ability to efficiently cr…