I need to train and evaluate the ConvTasNet model on speech enhancement for a custom dataset.
How should I go about formatting the data, and what changes need to be made in the run.sh file.