hoangtnm / hackathon

E.D.I.T.H Project by The Unsittables

Prerequisites

• Python 3.7+
• Ubuntu 18.04 LTS
• Docker Engine - Ubuntu (Community)
• Docker Desktop for Windows
• PyTorch
• Django
• OpenCV

Docker Installation Guide (Ubuntu)

sudo apt remove docker docker-engine docker.io containerd runc
sudo apt update
sudo apt install \
    apt-transport-https \
    ca-certificates \
    curl \
    gnupg-agent \
    software-properties-common
curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo apt-key add -
sudo add-apt-repository \
   "deb [arch=amd64] https://download.docker.com/linux/ubuntu \
   $(lsb_release -cs) \
   stable"
sudo apt update
sudo apt install docker-ce docker-ce-cli containerd.io
sudo usermod -aG docker $USER

Dataset structure

$ tree --dirsfirst --filelimit 10 Cyclone_Wildfire_Flood_Earthquake_Database
Cyclone_Wildfire_Flood_Earthquake_Database
├── Cyclone [928 entries]
├── Earthquake [1350 entries]
├── Flood [1073 entries]
└── Wildfire [1077 entries]
 
4 directories

Project structure

Our project contains:

• The natural disaster dataset.
• An checkpoint/ directory where our model and plots will be stored. The results from my experiment are included.
• A selection of videos/ for testing the video classification prediction script.
• Our training script, training.py . This script will perform fine-tuning on a ResNet18 model pre-trained on the ImageNet dataset.
• Our video classification prediction script, predict_camera.py , which performs a rolling average prediction to classify the video in real-time.

Run E.D.I.T.H

python manage.py runserver