A simple seq2seq spelling corrector which uses seq2seq based neural network to predict the correct spelling of an incorrect word.The inspiration behind this project was to see the powers of seq2seq.Below is the Neural Net graph for the same.
Original Graph:
Amplified Graph:
The dataset is taken from here and here then self preprocessed to convert all the files in one format and is present by the name final_dataset.csv.
1. Run main.py to train the model.
2. tensorboard --logdir=/tmp/SpellCorrector