load_admet

ADMET — pharmacokinetic and toxicity properties. Deprecated: this function calls load_dataset and will be removed.

Use instead

from torch_molecule.datasets import load_dataset

data = load_dataset("admet_ai")

Reference

load_admet(local_dir='torchmol_data', target_cols=None, return_local_data_path=False)[source]

Deprecated: use load_dataset("admet_ai").

Parameters:

  • local_dir (str, default="torchmol_data") – Directory used to cache downloaded files. Existing files are reused.

  • target_cols (list of str, optional) – Endpoints to select; default all 41 endpoints.

  • return_local_data_path (bool, default=False) – Also return the cache directory.

Returns:

  • dataset (SMILESDataset) – data is a list of canonical SMILES. target is a NumPy array of shape (n_molecules, n_targets), with NaN for missing measurements.

  • local_dir (str, optional) – The cache directory, returned alongside dataset only when return_local_data_path=True.

Return type:

SMILESDataset

See Datasets for caching, custom data, and train/test splits.