Project carried out for the course 02456 Deep Learning at the Technical University of Denmark by Peter Mørch Groth.
The main RMDN is trained in the main.ipynb-notebook. The model is contained in the model.py-script. The baseline GMM and MDN are trained and tested in the baselines.ipynb-notebook. Various visualisations are created in the visualisations.ipynb-notebook which uses functions from the util.py-script. Data pre-processing is carried out in - you guessed it - the data_preprocessing.ipynb-notebook. The saved models and their training and validation log-likelihoods can be found in the models folder.