/data has two files alice.en and alice.x. The later is created by applying a simple substitution
set of rules for each letter. Lets call X is the language after coding.
Task:
- Take 2000 first sentences of both files as training data, the next 200 sentences as development data and the rest for testing.
- Build a seq2seq model that learn to map english sentences to language X.
- Adding attention mechanism to the model, observing the performance with respect to the length of the input.
- Test the system on your testing data.