mental.py: the core model.train.py: the training script.extract.py: the audio feature extraction script.combine.py: the script to combine prediction and generate submission.
See solution.pdf for details. The code is copied from a working direction and needs to be modified to be runnable. Apply extract.py to preprocess the audio file and generate the features in pickle format. The training script does not save models, but instead saves the predictions. The combine.py script combines the predictions and generates the submission file.