Forked from https://github.com/masakhane-io/masakhane-ner

Data phonemized with epitran, for experiment.

• data/swa is as usual.
• data/swa_no_word_boundaries is edited to have character-level labels
• data/swa_phonemes we used epitran on each word of data/swa.
• data/swa_phonemes_no_word_boundaries we took above, and edited it to have character-level labels.

Did the same for Kinyarwanda:

• data/kin
• data/kin_no_word_boundaries
• data/kin_phonemes
• data/kin_phonemes_no_word_boundaries