*except downgrading to a later version from github. I am not entirely sure, how to do that yet ... might need to try that next, but I am literally on this training task for days now, not getting it to run on any machine. I did try later version on windows before, but failed unfortunately aswell there. I could also try my mac aswell, but I fear I might not be technically experienced enough for that one aswell (in that regard).
Shavkat Sultanov schrieb am Mittwoch, 7. Januar 2026 um 08:17:40 UTC+1: > Hi again, > > > I wouldn't know, how this might help, but here is my (small just to test > functionality) dataset: > > https://github.com/shavkat2610/gg_custom_1-ground-truth > > I did everything exactly as it is said in your manual for tesseract 5.x.x . > this one: > https://tesseract-ocr.github.io/tessdoc/tess5/TrainingTesseract-5.html > > Please help. > > > Dear regards, > Shavkat Sultanov > > Shavkat Sultanov schrieb am Dienstag, 6. Januar 2026 um 21:15:00 UTC+1: > >> Hi there, >> >> >> thanks in advance. >> >> this is what happens, when I try running the training script: >> >> [image: running_train_script.png] >> it is apparently failing to read the traineddata-file >> >> I downloaded it from here: >> >> https://github.com/tesseract-ocr/tessdata_best/blob/main/eng.traineddata >> >> my exact command to run it would be: >> >> sudo make training RATIO_TRAIN=1.0 MODEL_NAME=gg_custom_1 DATA_DIR=./data >> GROUND_TRUTH_DIR=./data/gg_custom_1-ground-truth START_MODEL=eng >> TESSDATA=/usr/local/share/tessdata MAX_ITERATIONS=500 >> >> >> [image: Screenshot 2026-01-06 185705.png] >> as you can see from this image, I am using tesseract version 5.5.2 >> >> my computer is running Ubuntu 24.04.3 but I tried windows before, failed >> aswell, but a little further down the process ... . >> [image: Screenshot 2026-01-06 190626.png] >> >> I have no clue why this is happenning. I would really like this to work >> though, because I have a particular problem, that is very monotone (easy, >> reading numbers off the screen, with the same font), but not being solved >> by the original eng.traineddata - model . >> >> Please help! I can provide additional info, if you ask me. >> >> Thanks in advance! again. >> >> >> Kind regards, >> Shavkat Sultanov >> > -- 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 visit https://groups.google.com/d/msgid/tesseract-ocr/6d91a2a0-6df9-4f05-8c0f-e71de62885f8n%40googlegroups.com.
