This traineddata file is fine-tuned using spanish trainneddata "https://github.com/tesseract-ocr/tessdata_best/blob/master/spa.traineddata" which was confusing slashed zeros with 8 or 9 on my dataset. This might be usefull if you are using tesseract and need to apply OCR on a dataset with slashed zeros. I used tesseract 4.0.0 beta.3 for training.