Skip to Content
ChemistryMolecular Generation

Molecular Generation

Generative models for drug and material design — VAEs, RL-based, diffusion, and graph-based approaches.

Tools Datasets Awesome Lists

Tools

Generative Models

REINVENT Reinforcement learning for molecular design GraphINVENT Graph-based molecular generation hgraph2graph Hierarchical molecular graph generation stk Building, manipulating, and analyzing molecules perses Expanded ensembles for chemical space exploration

Diffusion Models

GeoDiff Geometric diffusion for molecular conformation EDM Equivariant diffusion for 3D molecules DiffSBDD Structure-based design via diffusion

Benchmarks

GuacaMol Benchmarking for de novo molecular design MOSES Benchmarking platform for molecular generation

Virtual Screening & Docking

DiffDock Deep learning-based docking AutoDock Vina Popular open-source docking software AutoDock-GPU GPU-accelerated docking Gnina CNN-scoring molecular docking

Datasets

Chemical Libraries

ZINC20 Chemical library for deep docking virtual screening ZINC22 Commercially-available compounds for virtual screening GDB Enumerated molecules following chemical feasibility rules Enamine HTS 1.93 million diverse screening compounds

LLM & Generation Datasets

ZINC20-ML Deep-learning-ready ZINC20 formats 300M+ ChemPile Mixture-of-expert chemical corpus 75B+ tokens SmolInstruct Instruction dataset from 15 chemistry tasks 3.3M pairs

Awesome Lists

Molecular generation papers  Drug discovery papers  Molecular docking resources 
Property Prediction & ADMETQuantum Chemistry