How to fine-tuned paddle ocr to specific language like indonesian

[blizaga]

I managed to fine tune the paddleocr model but I used yaml configuration for English, what I want is paddle ocr can be multilingual, namely Indonesian with datasets that I collect myself.

And when I do inference, paddleocr only detects that it is English not Indonesian.

[GreatV]

please refer to: https://paddlepaddle.github.io/PaddleOCR/latest/ppocr/model_train/training.html

[blizaga]

I am still confused from the documentation, which part refers so that I can train paddleocr in Indonesian?