Appendix talk:Lithuanian pronunciation

Question: Would the community here welcome a mass-update of Lithuanian pronunciations, made consistent in a reasonably conservative way? Context: There is an ongoing SIGMORPHON shared task that involves Wiktionary-derived web-scrapes of Lithuanian word pronunciations, and those of 9 other languages too. For some of us attempting the task, it quickly became evident that there is quite a bit of inconsistency among the pronunciations. Another participant and I have outlined them in this GitHub issue. Some of it is not exactly the most unwieldy kind of inconsistency for humans, but I would reckon that for (1) those learning Lithuanian and (2) those interested in analyzing it linguistically, it might as well be. (As for machine models, consistent data is always better, but serving as a ML dataset is probably not a priority for Wiktionary.) - Ignas Rudaitis (talk) 06:42, 30 March 2020 (UTC)