This work is modeled after the work in Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification
To get ready for this lesson please download the data from here. Extracrt the two folders into the same directory as the jupyter notebook.
Then install the conda environment from the environment.yml
file by running conda env create -f environment.yml
And finally, if you haven't already, please go ahead and pull this repo. We'll be working within the jupyter notebook.
The first set of data come from the Kaggle gender-detection-face dataset. The second set of data comes from UTK: large scale face dataset Code is attributed to where I snagged it from, but the meat of the model training comes from this tutorial