Regarding fine-tuned traindata model size

39 views
Skip to first unread message

Nikhil Fande

unread,
Oct 18, 2021, 11:43:02 AM10/18/21
to tesseract-ocr
Hi @stweil , 
at the time of training, I have added model_name.wordlist, model_name.punc, model_name.numbers in data/model_name directory. the traindata resulted as 15mb file,

after unpacking the ~15mb traindata(my fine-tuned model) and ~22mb traindata( original traindata ), i found there are some files missing in fine-tuned model, so is it really necessary to add all those missing files in my fine-tuned model before inference ?

please find attachment of extracted folder structure of fine-tuned model and original traindata.
15mb_traindata.png
22mb_traindata.png
Reply all
Reply to author
Forward
0 new messages