Open zloidemon opened 6 years ago
this is related to #7 and #8 -- there isn't any Russian g2p model for sequitur available (AFAIK). There is a model for phonetisaurus somewhere on the internet (not sure where, @hpkit would know)
Hi, @jtrmal . I'm newbee with Kaldi. As I understand I may skipp step Preparing pronunciations for OOV words. My final model will work, but I get loewr accuracy becouse my model will not recognize out of vacabulary words. Am I right?