train.py errors when trying to train the ocr for using correctly openalpr tool

[mihaita....@gmail.com]

Hello all,

     I m trying to use openalpr for correctly identifying german plate numbers. I followed all the steps here :https://github.com/openalpr/train-ocr and I am trying to run the train.py script now.

     The platform is pi 3 and the OS Raspbian.

But I get a lot of errors. I investigated on internet about these type of errors and saw that several other people reported them and that through the suggested solutions was reinstalling teressact with training tools option enabled. 

    I have checked I have the source code for the training tools in tesseract/training directory and compiled them again. I have around 50 images  for german plates but I can't get pass this step with

train.py failing with the attached errors.

     Could somebody please give a suggestion about the possible reason for this issues or what and where should I read about it?

     Thanks in advance.

Best regards,

         Mihaita