Implementation of H-Transformer-1D, Hierarchical Attention for Sequence Learning - View it on GitHub