Hi, I met the same problem. Have you found the reason or solutions? Thanks 在 2019年7月16日星期二 UTC+8上午3:00:37，Artur Maricato Curinga写道： > > I'm trying to generate a model with OCR-B font to read documents MRZ, > tried an overfitting test using 44 cropped characters and trained using the > ocrd-training repo (https://github.com/OCR-D/ocrd-train) using train = > eval img list or without the eval on training. > > The training stops since lstmeval gives a 0% error rate, but testing the > model with tesseract fails to predict the correct text on 15% of the images. > What is the difference between the lstmeval and the tesseract prediction? > This behaviour is expected? >
-- 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 view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/123ee6c3-cc66-4a76-9816-629b75ed639e%40googlegroups.com.
