MolecularDiffusion.cli.predict¶

Prediction command for MolCraft CLI.

Adapted from scripts/predict.py for package-level execution.

Attributes¶

Functions¶

`count_atoms_from_xyz`(→ int)	Fast atom counter for XYZ files.
`is_rank_zero`()	Check if current process is rank zero.
`load_model`(chkpt_path[, task_config, atom_vocab])	Load a pre-trained model from checkpoint with auto-detection.
`log_hyperparameters`(object_dict)	Log hyperparameters for debugging.
`predict_main`(cfg)	Entry point for CLI predict command.
`runner`(→ Tuple[Dict[str, Any], Dict[str, Any]])	Property prediction run.
`xyz2mol`(xyz_file, atom_vocab, node_feature[, ...])	Convert an XYZ file into a PyTorch Geometric Data object.

Module Contents¶

MolecularDiffusion.cli.predict.count_atoms_from_xyz(path: str) → int¶: Fast atom counter for XYZ files.

MolecularDiffusion.cli.predict.is_rank_zero()¶: Check if current process is rank zero.

MolecularDiffusion.cli.predict.load_model(chkpt_path, task_config=None, atom_vocab=None)¶: Load a pre-trained model from checkpoint with auto-detection.

MolecularDiffusion.cli.predict.log_hyperparameters(object_dict: dict)¶: Log hyperparameters for debugging.

MolecularDiffusion.cli.predict.predict_main(cfg: omegaconf.DictConfig)¶: Entry point for CLI predict command.

MolecularDiffusion.cli.predict.runner(cfg: omegaconf.DictConfig) → Tuple[Dict[str, Any], Dict[str, Any]]¶: Property prediction run.

MolecularDiffusion.cli.predict.xyz2mol(xyz_file, atom_vocab, node_feature, edge_type='fully_connected', radius=4.0, n_neigh=5, device='cpu')¶: Convert an XYZ file into a PyTorch Geometric Data object.

MolecularDiffusion.cli.predict.log¶