materials.mhg-ged

Introduction

We present MHG-GNN, an autoencoder architecture that has an encoder based on GNN and a decoder based on a sequential model with MHG. Since the encoder is a GNN variant, MHG-GNN can accept any molecule as input, and demonstrate high predictive performance on molecular graph data. In addition, the decoder inherits the theoretical guarantee of MHG on always generating a structurally valid molecule as output.

Table of Contents

1. Getting Started
   1. Pretrained Models and Training Logs
   2. Installation
2. Feature Extraction

Getting Started

This code and environment have been tested on Intel E5-2667 CPUs at 3.30GHz and NVIDIA A100 Tensor Core GPUs.

Pretrained Models and Training Logs

We provide checkpoints of the MHG-GNN model pre-trained on a dataset of ~1.34M molecules curated from PubChem. (later) For model weights: HuggingFace Link

Add the MHG-GNN pre-trained weights.pt to the models/ directory according to your needs.

Installation

We recommend to create a virtual environment. For example:

python3 -m venv .venv
. .venv/bin/activate

Type the following command once the virtual environment is activated:

git clone git@github.ibm.com:CMD-TRL/mhg-gnn.git
cd ./mhg-gnn
pip install .

Feature Extraction

The example notebook mhg-gnn_encoder_decoder_example.ipynb contains code to load checkpoint files and use the pre-trained model for encoder and decoder tasks.

To load mhg-gnn, you can simply use:

import torch
import load

model = load.load()

To encode SMILES into embeddings, you can use:

with torch.no_grad():
    repr = model.encode(["CCO", "O=C=O", "OC(=O)c1ccccc1C(=O)O"])

For decoder, you can use the function, so you can return from embeddings to SMILES strings:

orig = model.decode(repr)

For more information contact indra.ipd@ibm.com