TunedM2M

This repository hosts code for the paper Combining Transfer Learning And Transformer Attention Mechanism to Increase Aqueous Solubility Prediction Performance.

Installation

Prequisites

*Python 3.7

Setup

git clone https://github.com/magdalenawi/TunedM2M
cd TunedM2M
pip install -r requirements.txt
conda install rdkit -c rdkit
conda install pytorch=0.4.1 torchvision -c pytorch
pip install torchtext==0.3.1
pip install dgl

Datasets

Available here.

Experiments

For the experiments one can use the datasets found in the link above. In the simplest scenario, you only need to run the predict_demo.py script or a Jupyter notebook called run_demo.

To do

project files and demo 🎉
add more demo files (with graphics)
add docs
...