== This is an implmentation of Encoder-Decoder LSTM model ==

1. add Vanilla Encoder-Decoder without Attention -- congratulations! this version can be used!

==v2.0== This version I will mainly add NER support

1. add pre-processing code

==v2.5== support UNK replacement

==v2.6== support shuffle per epoch

==v2.6.5==

1. merge encdec-attention-ner-unk-margin-v3.5 on -- view the real output -- 28th Feb: add bleu evaluation; using mteval-v13a.pl -- 28th Feb: view output -- 29th Feb: add FKGL evaluation; using https://github.com/mmautner/readability -- 29th Feb: A whole pipeline from loading model to get the final evaluation results and analysis file

==v2.6.6==

1. add word embedding support (should be fast to implement)
2. add newela preprocessing code

==v2.6.10==

1. add EncDecAWE.lua -- fineTune works this time