Getting Started

Installation

Set up conda environment and clone the github repo

# create a new environment
$ conda create --name synthmol python==3.8.10
$ conda activate synthmol

# install requirements
$ pip install torch==2.0.1
$ pip install torch-geometric==1.7.2
$ pip install numpy==1.22.1
$ pip install pandas==2.0.3
$ pip install rdkit==2023.9.4
# Uni-Mol install https://github.com/dptech-corp/Uni-Mol.git

# clone the source code of SynthMol
$ git clone https://github.com/ThomasSu1/SynthMol.git
$ cd SynthMol/src

Dataset

You can download the dataset file under SynthMol/Data/ folder.

Training

To train the SynthMol model, where the configurations and detailed explanation for each variable can be found in model_train.py.

# Navigate to the source directory
$ cd SynthMol/src

# Run the training script
$ python model_train.py

Acknowledgement

• Uni-Mol: https://github.com/deepmodeling/Uni-Mol/tree/main