Inltk

1. English support has been added
2. Runtime performance of get_similar_sentences has been added

This contains addition of new feature - get_similar_sentences - with which you can augment and multiply your data in supported languages

New Features:

* You can now get 400 dimensional encoding for sentences using `get_sentence_encoding` - supported for all languages in iNLTK
* You can now get similarity score (cosine similarity) between 2 sentences using `get_sentence_similarity` - supported for all languages in iNLTK.

New Model:
* The above features will not work for *punjabi language* with the old model. Please execute the following code-snippet before using them

from inltk.inltk import reset_models

>> reset_models('pa')
>> setup('pa')

Added Urdu support to iNLTK - thanks to anuragshas contributions
Added Windows 10 support - thanks to ibrahiminfinite contributions

Added get_embedding_vectors function to allow users to get embedding vectors for their words/sentences/documents

Added tamil support