The files contained in this repository are part of the thesis "Filtered Feelings: Investigating Frequency Filters in Speech Emotion Recognition Models".
The IPython Notebooks contain all the code to perform the experiments as they were done in the thesis. If you plan on replicating this research with the same model, make sure to download it from here. The notebooks assume that the model files are in a folder "speechbrain" in the same folder as the notebooks.
The CSV files in the "results" folder are all the CSV files that were generated for the thesis.
In this research the IEMOCAP and MELD datasets were used.
The IEMOCAP can only be downloaded once you request access to it.
The MELD dataset can be downloaded on their GitHub page here.