Official implementation of Large Language Model Based Multiple Property Prompts Learning Framework for Molecular Property Prediction

⏳ Quick Start

1. Installation

   pip install -r requirements.txt

Furthermore, you also need to download the pre-trained weights of SCI-BERT. For the molecule edit task, you should install MolSTM first.

2. Pretraining

   python3 -u moltex_train.py

To start the pre-training stage, you can run our code with above command.

3. Property Prediction on PharmaBench/MoleculeNet

   python3 -u ft_pha.py
   python3 -u ft_mol.py

To fine-tune ont the PharmaBench and MoleculeNet datasets and test on them, you can run our code with above commands.

4. Molecule Edit Task

After installing all packages, you can directly run the .ipynb files with Jupyter to run this task.