The purpose to create this project is if we give input as wrong sentence then it will automatically correct and will give an output as right sentence.
I have taken data from http://www.gutenberg.org/ebooks/search/?sort_order=downloads which is composed of twenty popular books.
This model is designed using grid search to find the optimal architecture, and hyperparameter values. The best results, as measured by sequence loss with 15% of our data were created using a two-layered network with a bi-direction RNN in the encoding layer and Bahdanau Attention in the decoding layer.
#Command to get output on Tensorboard tensorboard --logdir=path/to/log-directory