Dual_Stage_Attention_RNN

DARNN

Paper : https://arxiv.org/abs/1704.02971

Requirements

• python 3.7
• Tensorflow 2.3
• numpy
• pandas

Dataset

Dataset : https://dacon.io/competitions/official/235584/overview/

Data preprocessing : preprocess.py

Preproeceesing data : https://drive.google.com/drive/folders/13LnUp21gizx5IRFyN0ExrepHZajsUImz?usp=sharing

Data shape : inputs : [enc_data , dec_data] enc_data shape : batch,T,n dec_data shape: batch,T-1,1

implementation : Dual_stage_attention_model.py

Architecture

Result