Hello, I just finished my first training of tesseract 4.0 and I ran a lstmeval on the generated model, which I named *mod01.* I use this command line : lstmeval --model data/checkpoints/mod01_checkpoint --traineddata ./usr/share /tessdata/mod01.traineddata --eval_listfile data/list.eval
It worked fine and it gave me a character error rate and a word error rate. Now I would like to know if my training improved Tesseract's accuracy on my specific documents. So I wanted to launch the evaluation on the same dataset but with the model I started the training from, the english provided on Tesseract's github repo : eng.traineddata. I tried : lstmeval --traineddata ./usr/share/tessdata/eng.traineddata --eval_listfile data/list.eval But it did not work because I did not provided any --model And this showed me that my understanding of Tesseract's was not correct. Since downloading a new *lang.traineddata* is enough to use Tesseract with this lang I thought that all the model was contained in the traineddata files. What is this --model argument then ? In which my research on the web told me to put the last checkpoint of my training but without explaining why. Is it possible then to run lstmeval on a pretrained model like eng.traineddata ? Thank you ! -- 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 post to this group, send email to tesseract-ocr@googlegroups.com. 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/5f762b56-f7b0-4438-a8cb-cbab94304341%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
