CLI to create a pronunciation dictionary by predicting English ARPAbet phonemes using seq2seq model from g2pE and the possibility of ignoring punctuation and splitting on hyphens before prediction.