Interpreting the outputs of LSTM/GRU for the text classification task using modified bahdanau attention mechanism
What does this project solve
If datascientists want to build an interpretable text classification model, they can use this model because it has inherent interpretation capabilities.
- Developed a modified bahdanau attention model to interpret the LSTM/GRU models for text classification (Both binary and multi-label)
- Dataset : IMDB review dataset