Access to source materials?

[SeyelentEco]

I'm working on a project that requires similar data in order to build models to aim with improving readability of older Hawaiian language texts, do you still have access to the source data? I am working on building the dataset myself but I stumbled upon your data and I thought it might save me some steps.

[bee-san]

Hi @SeyelentEco , I cannot help you with source data but to necro this...

Did you build anything for older hawaiian texts? I am also very interested in this and curious as to what you did :)

[SeyelentEco]

Yes I used existing text from Wehewehe and removed 'okina and kahakos from existing text and then fine-tuned a ChatGPT4 model to learn how to add them back. You can play with it here: https://olelotools.com/