Code runs on Python 3.7.8 (Clang 11.0.3)

Use pip_requirements.txt to load required libraries:

pip install -r pip_requirements.txt

NLTK stopwords and scraped Hatebase terms are included in the project.

To install word embeddings files, uncomment lines 173-175 in ve/main.py and run. Uncompress downloads and copy files to ve/Vectors folder.

enchant library is include in pip_requirements.txt but underlying library should be installed on OS:

e.g on Mac OS:

brew install libomp
install brew update & brew install enchant

See https://pyenchant.github.io/pyenchant/install.html for other examples.

Dataset should be installed as a tab-seperated file in ve folder with with the name founta-dataset.csv.

Code expects a header line. If necessary, tabs should be escaped with "

Example first lines with the label hateful:

text	label confidence
This is some text	hateful 4

Run python /ve/main.py to run all models