Projects using this and evaluation results

[NebelAI]

Hi @kwonmha,

your project is exactly what came into my mind when dealing with Bert vocab creation. Currently I'm doing some vocab optimizations for my Bert project, too.

Can you say something about improvements/degradations related to your vocab changes? I'm really curious if this approach delivers better results.

[kwonmha]

Well, I haven't trained BERT for many times with different vocab types.
This is the only vocab I tried that has the same format with official google research's BERT.
So there's nothing to compare.

I have plans to utilize pos tag info with subwords as I'm doing research on Korean.
But I'm not sure it will work on English or other alphabet-based languages.

[YuBeomGon]

Can I know the input file pattern for korean text?