Work with drug-discovery datasets and benchmarks.
Loads therapeutic machine-learning datasets, prepares suitable splits, evaluates predictions, and scores molecules through PyTDC.
When to use it
Use when working with TDC therapeutic machine-learning datasets or benchmarks.
Give it a therapeutic-data request; it returns bounded discovery, split, evaluation, or molecular-scoring summaries, confirming download scope before execution.
Requires uv for the documented environment setup and CLI invocations.
Uses CPython 3.11 as the reproducible runtime target because the dependency constraints exclude newer interpreters.
This skill
Harvard Dataverse
Downloads approved datasets and benchmarks
TDC task pages
Checks dataset licenses and terms