A powerful and flexible machine learning platform for drug discovery

TorchDrug is a machine learning platform designed for drug discovery, covering techniques from graph machine learning (graph neural networks, geometric deep learning & knowledge graphs), deep generative models to reinforcement learning. It provides a comprehensive and flexible interface to support rapid prototyping of drug discovery models in PyTorch.

$ conda install -c milagraph -c conda-forge torchdrug
A powerful and flexible machine learning platform for drug discovery

Key Features

Minimal Domain Knowledge

Build and train machine learning models for drug discovery with minimal domain knowledge.

Datasets and Building Blocks

Empower fast iteration of ideas by a large collection of common datasets and building blocks.

Comprehensive Benchmarks

Benchmarks provide a systematic comparison of deep learning architectures for drug discovery.

Scalable Training and Inference

Seamlessly scale models to multiple CPUs, multiple GPUs, or even distributed settings.