Is there a step-by-step instruction for training the model for the Arabic language?

[AboulfazlSeilsepour]

I plan to train the current model on the local system and add Arabic to it. The documentation provides very general instructions on how to do it. How do you prepare the system for execution, what files are required in a specific format (need JSON template sample and explain limitations in size or formats) for training the model, what commands need to be executed one after the other, how do you save them, etc?
I downloaded GOT_weights files.

my system hardwares: 12700K-CPU -> 32GB Ram -> Nvidia RTX3050Ti
OS: Windows 11
checkout repositories: GOT-OCR2.0 / ms-swift / deepspeed