The HAN model is based on text classification project of Heet Sankesara: https://github.com/Hsankesara/DeepResearch/tree/master/Hierarchical_Attention_Network
Datasets are not included. glove embedding with 100 dimensions is used but not included in the repository.