# Dataset

Dataset in this submission is tokenized with [moses](https://github.com/moses-smt/mosesdecoder/tree/master/scripts/tokenizer), used: `tokenizer.perl` (with pl language) and `deescape-special-chars.perl`.