You want to do grid search. The best practice is to try each of those parameters over a range and try all combinations. If you find that your results are improving at the one or another extreme rather in the middle of the range, you'll want to pick a new set of parameters overlapping your margins.
On Fri, Feb 9, 2024 at 7:49 AM Omar Samir <omaredu1...@gmail.com> wrote: > Now I am trying to fine tune Tesseract Arabic model in the testdata_best > with more than 20K images in the dataset. I want to know the best values > for parameters like MAX_ITERATIONS, EPOCHS, and LEARNING_RATE so how can I > find these values I also look inside research papers but I find nothing > about the best values for training a model. > > -- > 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 tesseract-ocr+unsubscr...@googlegroups.com. > To view this discussion on the web visit > https://groups.google.com/d/msgid/tesseract-ocr/25365e5b-f591-48d7-8819-124e87be8c0an%40googlegroups.com > <https://groups.google.com/d/msgid/tesseract-ocr/25365e5b-f591-48d7-8819-124e87be8c0an%40googlegroups.com?utm_medium=email&utm_source=footer> > . > -- 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 tesseract-ocr+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/CAL7mBq56m1uG2D25hyD%3DmeAK%2B3pOaRE6hy-L2xD6SS0Uvbe%3D2w%40mail.gmail.com.
