what is the size of fine tuned traineddata

21 views
Skip to first unread message

Shanshan Wang

unread,
Apr 23, 2019, 12:24:34 PM4/23/19
to tesseract-ocr
Hi,

I used the ocrd_train for fine tune training, the start model is eng.traineddata in tessdata_best. 
the traineddata file I got after training is around 11-12MB, less than the original eng.traineddata which is 15MB.

Is it normal? or I have done something wrong during training?  I have added unicharset, punc, wordlist and dictionary during training. 

Thanks,
Reply all
Reply to author
Forward
0 new messages