Robert Sachunsky
Robert Sachunsky
Agreed. Can we get something similar in core based on OtoN, though?
I can confirm this is still an issue. In my case, the difference is much worse: | plot_log | plot_cer | | --- | --- | |  is the same, only the extraction method differs. Anyway,...
@zdenop I am trying to understand. So the GT text file is bogus, but not to the extent that generate_line_box.py would raise an exception and thus cause non-zero exit value,...
> Meanwhile, it turned out that the error values estimated during the training process are not to be trusted: [#261](https://github.com/tesseract-ocr/tesseract/issues/3560) This has been mitigated to some extent by https://github.com/tesseract-ocr/tesseract/pull/3644, which...
IMO this should be done based on the recipe [described for the API changes](https://github.com/OCR-D/core/issues/322#issuecomment-687370354). So this would require a rewrite of all bashlib processors – but we need to do...
> `add` also created this structMap: that's also what I witnessed as prime problem in https://github.com/OCR-D/ocrd_tesserocr/issues/201. We need more diagnostics why and exactly when this is happening. But this is...
> I'll open another GitHub issue for that, if you didn't already. No, please do!
> * Changing the workflow to run requires updating core ok, good point! > * The reusable parts, i.e. up RabbitMQ, MongoDB and processing server is generic enough that the...
Probably a problem with recent TensorFlow updates. We already migrated to 2.0, but not sure which was the last tested version. Could you try replacing `s[1]` with `s[1] or -1`...