This directory contains example scripts for OCR on the Yomdle and Slam datasets.
  Training is done on the Yomdle dataset and testing is done on Slam.
  LM rescoring is also done with extra corpus data obtained from various sources