Loss Functions

A loss term that penalizes predicting cortical labels that aren't adjacent to the true label. Implemented in PyTorch and DGL and used as a regularizer alongside standard cross-entropy.