MolecularDiffusion.modules.models.nextmol¶
NExT-Mol: 3D diffusion (DMT) meets 1D language modeling (MoLlama).
Liu*, Luo*, Huang et al., NExT-Mol: 3D Diffusion Meets 1D Language Modeling for 3D Molecule Generation, ICLR 2025 (arXiv:2502.12638).
The two halves are fully decoupled – their only interface is a flat list of
SMILES. mollama writes the molecule down; dgt places it in 3D.