rm -f eng.*
for f in cube.{bigrams,fold,lm,nn,params,size,word-freq} tesseract_cube.nn traineddata; do
f=eng.$f
- termux_download https://raw.githubusercontent.com/tesseract-ocr/tessdata/master/$f $f
+ # From the tessdata README: "These language data files only work with
+ # Tesseract 4. They are based on the sources in tesseract-ocr/langdata on GitHub.
+ # Get language data files for Tesseract 3.04 or 3.05 from the 3.04 tree."
+ termux_download https://raw.githubusercontent.com/tesseract-ocr/tessdata/3.04.00/$f $f
done
}