Generate novel crystal structures with multimodal property conditioning using a reciprocal-space DiT. Returns CIF data.
Multimodal DiT for inorganic crystal generation. Conditions on composition, crystal system, space group, band gap, formation energy, e-above-hull, and magnetic ordering.
Generate novel crystal structures for a given chemical composition using a diffusion transformer with periodic representation, AdaLN conditioning, density-based element classification, and learned lattice prediction. Returns CIF file data with structural metrics.
Diffusion transformer with periodic fractional coordinate representation and AdaLN composition conditioning. Operates on 128-cubed voxel density fields with learned lattice prediction.
Generate novel crystal structures for a given chemical composition using a 1.2B parameter diffusion transformer. Returns CIF file data with structural metrics.
Diffusion transformer operating on 128 cubed voxel density fields.