This repository contains the final project for a course in parallel programming, focusing on distributed model training using the Message Passing Interface (MPI). The project extends the DeepPurpose framework developed by kexinhuang12345, integrating multi-node capabilities to enhance drug-target interaction (DTI) prediction models.
In our project, we implemented the mpi_train
function in the DTI.py
script of the DeepPurpose framework to enable distributed training across multiple nodes. This approach aims to improve the training efficiency and scalability of the DTI models.
- Midterm Proposal: Our initial project proposal, outlining the project goals and preliminary methods, is available here.
- Final Paper Report: Our comprehensive report, titled "Distributed Model Training with MPI for Multi-Node Training," details the methodologies, results, and conclusions of our distributed training implementation. It can be accessed here.