Combining traineddata files

Umanda Dikwatta

unread,

Oct 22, 2022, 2:09:19 AM10/22/22

to tesseract-ocr

Hello,

I have several traineddata files for different fonts for the same language. How to combine these files to get one traineddata file ?

Best Regards

Zdenko Podobny

unread,

Oct 22, 2022, 3:30:47 AM10/22/22

to tesser...@googlegroups.com

Legacy or LSTM?

Zdenko

so 22. 10. 2022 o 8:09 Umanda Dikwatta <abey...@gmail.com> napísal(a):

Hello,

I have several traineddata files for different fonts for the same language. How to combine these files to get one traineddata file ?

Best Regards

--
You received this message because you are subscribed to the Google Groups "tesseract-ocr" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tesseract-oc...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/3d35b7d7-636b-42b4-a5b9-9bfb8951c234n%40googlegroups.com.

Umanda Dikwatta

unread,

Oct 22, 2022, 3:36:24 AM10/22/22

to tesseract-ocr

LSTM

Zdenko Podobny

unread,

Oct 31, 2022, 5:27:10 AM10/31/22

to tesser...@googlegroups.com

As far as I know, there is no tool that can merge finished training.

IMO for LSMT it should work as incremental training e.g. you used the output of fine-tuning as input for the next fine-tuning... (I never tried it ;-) )

Zdenko

so 22. 10. 2022 o 9:36 Umanda Dikwatta <abey...@gmail.com> napísal(a):

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/224afc45-6461-4c1c-a3e6-c60d312f2b4an%40googlegroups.com.

Reply all

Reply to author

Forward