This repository contains all the code used to fine-tune GPT-3 models for the prediction of antimicrobial activity and hemolysis of peptide sequences as described in the preprint "GPT-3 accurately predicts antimicrobial peptide activity and hemolysis". Please cite our work if you use any of the code included in this repository. Mind that this repository is a work in progress and subject to regular updates.
The repository contains:
- data directory: Contains all unprocessed and processed data used to fine-tune the GPT-3 models.
- Jupyter notebooks: Each notebook contains all the code necessary to process the data and fine-tune the GPT-3 models.
The code can be run directly from the provided Jupyter notebooks. Before you start, make sure to install all required packages in your virtual environment. Code cells that are commented out in the notebooks are to be pasted in the terminal (while being in the project directory).
