This is the code that was submitted together with the paper at ISMIR 2024.
We strongly advise to set up a virtual experiments for these experiments.
pip install -r requirements.txt
You will also need to download additional resources from nltk in Python in your virtual environment.
nltk.download('stopwords')
nltk.download('punkt')
nltk.download('wordnet')
nltk.download('averaged_perceptron_tagger')
nltk.download('vader_lexicon')
For more clarity, we describe the different scripts to run to reproduce our experiments in a separate README.
TO-DO-G: add the notebook in src/models
Below an overview of the main content of the code, that is in the src
folder:
configs
: configuration.yaml
files for the regression layersdata_prep
: all scripts related to data preparation for model trainingmodels
: all modelsembeddings.py
: extract embeddings from a modelfeatures.py
: stylometric featureshelpers.py
: generic helpers