Code to support the Asgard low powered computer vision cluster at Loyola University Chicago
Asgard is a simple cluster of NVIDIA Jetson Nano 2GB devices that aims to provide a test bed for running and testing Computer Vision models for low powered computing devices.
This repository contains all of the information regarding:
- Setting up a NVIDIA Jetson 2GB for this project
- The frontend web interface for uploading jobs (
.imgfiles) to the head node of the cluster - The backend server code to handle file transmission and storage
- Scripts to send files from the head node to a node
- Architecture diagrams to explain how this cluster is organized
- File format documentation to explain what is an isn't a valid file format
See JETSON_SETUP.md for more information.
Frontend code can be found here.
The frontend website is written in a single HTML, CSS, and JS file. It uses Pico.css as its CSS framework to simplify development.
The server backend for the website is handled via Flask and Python 3.10+.
Backend code can be found here.