About

This repository contains a convolutional neural network that is able to detect buildings on satellite imagery from SENTINEL-2.

Requirements

• We recommend a computer with a GPU
• at least 24 GB RAM and additionally 16 GB swap memory available
• Python 3.x installed on your system. You can download it from the official Python website.
• The connection to copernicus.eu requires a registration which must be done manually.

Setup Instructions

1. Clone this repository to your local machine

git clone https://github.com/christianstubbe/architecture-of-ml-systems
cd architecture-of-ml-systems

2. Create a virtual environment

python3 -m venv .venv
source .venv/bin/activate # On Windows, use `venv\Scripts\activate`

3. Install Dependencies

pip install -r requirements.txt 

Logging

Logs for this pipepline are stored in main.log.