Explore therapeutic data with safer, reproducible workflows.
therapeutic machine learning
Enables discovery, approved data access, splitting, evaluation, benchmarking, and bounded molecular-oracle workflows for therapeutic machine learning.
When to use it
Use when discovering therapeutic ML tasks, loading approved datasets, evaluating predictions, or working with benchmark groups.
Give it a therapeutic machine-learning request; it runs the applicable discovery, data, evaluation, benchmark, or oracle workflow with bounded project outputs.
This skill
remote/docking oracles
Sends molecules to oracle services
Harvard Dataverse
Downloads data from Harvard Dataverse
supported oracle checkpoints
Downloads supported oracle checkpoints
dataset task pages and source terms
Reads dataset licensing terms
Requires uv to create the environment and run the pinned packages.
Requires an isolated CPython 3.11 environment.
Requires the reviewed PyTDC 1.1.15 package snapshot.
Requires setuptools 80.9.0 because PyTDC imports the legacy pkg_resources module at runtime.