I want to make specific training data to use for sick leave forms and am trying to create the required files for it, but I am running into bugs.
After creating the box file and checking it I select the training method, "Train with Existing Box" or "Train from Scratch", the effect is the same, I always get this error:
The required file font_properties, frequent_word_list or word_list does not exist.
But all 3 files exist in the same directory as the tiff image and box file.
A second bug is that I cannot change the directory of the Tesseract executables, they are not at the path that was automatically chosen and the text cannot be changed, nor does the brows button actually do anything.
And even whe I created this directory and copied all of tesseract into it, I keep getting the above mentioned error.
I tried version 1.0 and the 2.0 Beta version but it is the exact same in both.
I don't know if it makes a difference but the language is Geran (deu), though I don't think I am even yet at the stage where language matters.
I'm on Windows 10 Pro build 10586, 64bit.
Last edit: First Last 2017-02-07
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Yes, the exact same thing happens in both v 1.0 and the v2.0 beta. Also, this is unrelated to not being able to change the location of the tesseract executables.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Sorry, we were unable to reproduce the issue. Since the training exe are bundled, you may consider following the Tesseract training wiki and doing it manually.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I want to make specific training data to use for sick leave forms and am trying to create the required files for it, but I am running into bugs.
After creating the box file and checking it I select the training method, "Train with Existing Box" or "Train from Scratch", the effect is the same, I always get this error:
But all 3 files exist in the same directory as the tiff image and box file.
A second bug is that I cannot change the directory of the Tesseract executables, they are not at the path that was automatically chosen and the text cannot be changed, nor does the brows button actually do anything.
And even whe I created this directory and copied all of tesseract into it, I keep getting the above mentioned error.
I tried version 1.0 and the 2.0 Beta version but it is the exact same in both.
I don't know if it makes a difference but the language is Geran (deu), though I don't think I am even yet at the stage where language matters.
I'm on Windows 10 Pro build 10586, 64bit.
Last edit: First Last 2017-02-07
Those files need to be prefixed with the language code (see
samples\viefolder for example).Tesseract executables points to the directory containing Tesseract executables. On Windows, it is preset to
tesseract-ocrsubdirectory.That is exactly what I did.
deu.font_properties
deu.frequent_word_list
deu.word_list
You have
deuin the Language textbox on Trainer tab and specify the location of the files in Training Data textbox?Last edit: Quan Nguyen 2017-02-08
yes
Have you tried on the bundled sample?
Yes, the exact same thing happens in both v 1.0 and the v2.0 beta. Also, this is unrelated to not being able to change the location of the tesseract executables.
Sorry, we were unable to reproduce the issue. Since the training exe are bundled, you may consider following the Tesseract training wiki and doing it manually.
Please try again with the latest versions, which have enhancements with regards to these required files.