This implementation is based on the SincAlignNet model from the paper 'Frequency-Based Alignment of EEG and Audio Signals Using Contrastive Learning and SincNet for Auditory Attention Detection'.