ACME Digit Classification

A machine learning framework for handwritten digit recognition, developed as part of the FYS-8805 Collaborative Coding course at UiT. Supports three datasets (MNIST, SVHN, USPS) via a unified CLI and Python API.

This document includes a User Guide for installation and CLI usage, an API Reference for the Python modules, and an Individual Contributions section describing each team member’s work.

Table of contents

• About this code
  • Getting started
  • Need help?
• Model development
  • Training
  • Evaluation
  • Testing
  • Contributing to the code
• Running on LUMI
  • Environment setup
  • Training
  • Evaluation
  • Monitoring jobs
  • Tips
• Individual contributions
  • Andrea Dahlmo Løkke
  • Keyne Oei
  • Myrthe Catteau
  • Rahul Baburajan
  • Riccardo Gelato
  • Siyan Chen
• API Reference
  • Data (src.data)
  • Models (src.models)
  • Training (src.training)
  • Evaluation (src.evaluation)
  • Inference (src.inference)