hypospadias-multi-task-learning

This project is focused on developing a multi-task-learning model to determine the HOPE and GMS scores for male infants with hypospadias using natural images.

This research is done as part of the Goldenberg Lab and SickKids Hospital.

Quick Setup

Go to project directory

# Go to `hypospadias-multi-task-learning` repo directory
cd PATH/TO/DIRECTORY

0. (Optional) Create virtual environmeent

# Create environment
python -m venv hypospadias_mtl

# Activate environment
# 1. In Windows
hypospadias_mtl\Scripts\activate
# 2. In Linux
source hypospadias_mtl/bin/activate

1. Install dependencies

pip install -r requirements.txt

2. Create symbolic link to data directory

# (In a console with admin priviliges)
# 1. In Windows
mklink /d ".\src\data" "PATH\TO\DATA\DIRECTORY"

# 2. In Linux
ln -s /PATH/TO/DATA/DIRECTORY ./src/data

3. Modify model training specifications in main.py

# Update model training specifications in main.py as needed to run experiments.