xnn.common.data.hub#

Dataset hub: HuggingFace load_dataset()-style one-liner loading.

Download and preprocess upstream atomistic datasets into xnn’ native structure dicts (optionally ready-to-train AtomicDataset objects) with a single call:

from xnn.common.data import load_dataset

splits = load_dataset("rmd17", molecule="aspirin")          # {"train", "test"}
train = load_dataset("rmd17", molecule="aspirin",
                     split="train", cutoff=5.0)             # AtomicDataset

Files are cached (and MD5-verified) under datasets/<name>/ in the repo by default; see default_cache_dir(). Add a new dataset by subclassing DatasetBuilder and calling register_dataset().

Modules

ani1

ANI-1 dataset builder (pyanitools HDF5 from figshare).

ani1ccx

ANI-1ccx dataset builder (coupled-cluster subset of the ANI-1x release).

ani1x

ANI-1x dataset builder (single pyanitools HDF5 from figshare).

ani2x

ANI-2x dataset builder (pyanitools HDF5 from the Zenodo release).

argon_md

Argon MD dataset builder (bundled with xnn).

base

Dataset registry and the load_dataset entry point.

lode_dimers

LODE non-bonded interactions dataset builder (molecular dimers & toy systems).

rmd17

Revised MD17 (rMD17) dataset builder.
Show Source