Produces: lat.traineddata
You need wget, unzip and the Tesseract training tools to make this training.
The following files have been automatically generated using the tools in the lattraining git repository located at https://github.com/ryanfb/latinocr-lattraining
You can see the exact process for generating them in the lattraining Makefile.
The Latin.unicharset file has been copied from Tesseract's tesseract-ocr.langdata git repository.