DeepChem
Democratizing deep learning for the sciences
DeepChem is a batteries-included library for applying machine learning to chemistry and biology. It bundles featurizers, benchmark datasets (MoleculeNet) and model implementations.
Highlights
- Featurizers: fingerprints, graph convolutions, 3D descriptors.
- Models: GNNs, random forests, transformers and more.
- MoleculeNet benchmarks for property prediction.