Hi,
I'm new in the Tesseract community and I recently submitted a PR [1].
Secondly, what if an existing traineddata set (X) was trained from another base (say, script A, which is alphabetic and left-to-right , being trained from Latin, which is also alphabetic and LTR), but community decided to train from scratch because the current model hallucinates due to the Latin base? Do they submit their newly trained data (Y) a new PR on tesseract-ocr/tessdata or tessdata_contrib? But how Y will be integerised since X was trained by fine-tuning Latin?
Thanks in advance!
Subhashish