Recurrent Neural Network(LSTM) for Sequence Prediction, here the encoder-decoder architecture is used to predict the output sequence.
For character wise seq-to-seq prediction:
run the code
python LSTM-seq-generator.py
For word wise seq-to-seq prediction:
run the code
python LSTM-words-generator.py
Hyperparameter | Value |
---|---|
batch_size | 10 |
epochs | 500 |
validation_split | 0.05 |
embedding_size | 50 |
LSTM layers | 1 |