Pretraining Graph Transformers with Atom-in-a-Molecule Quantum Properties for Improved ADMET Modeling

This repository contains the code to reproduce the results presented in this paper. The Graphormer model used here is the one from chytorch.

Table of Contents

• Installation
• Usage
• License

Installation

To install the necessary packages run the following command:

pip install -r requirements.txt

Usage

In order to run trainings and pretrainings use the following command:

python main.py ./yaml_files/yaml_file_of_choice.yaml

where 'yaml_file_of_choice.yaml' is a placeholder for the chosen file.

Curated datasets for pretraining are provided here, place them in a directory and change paths in the code accordingly. TDC datasets download and handling is done using PyTDC

For what concerns the analysis they can be run using the provided notebooks.

License

The code is provided under MIT license, for the curated datasets we refer to the licenses reported in the corresponding links.