Currently, Ray/Google has NOT released info on how to train Tesseract 4 (LSTM) with real life images. The only supported option is to use synthetic training data created by tesstrain.sh script using training text and unicode fonts.
To train an LSTM model from scratch requires a large amount of training data and huge computing resources and time (in days/weeks). As a user, your best bet for training is to try finetuning for a particular font or adding a couple of characters. -- 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 [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/tesseract-ocr. To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/CAG2NduWpKdwMKQBgNhFK9m_PEn4EZ9GsCZO5juwoPSVzE1dpKA%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
