Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.
And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.
the correspondence about tags or labels
See original GitHub issueHi,
I am quite confused about the tags in training files and resulting tei files in prediction phase. Tags used for annotation seems not the same to the tagging label, they have a mapping relationship but I fail to find that. Such as in the tagginglabels file, there is <section>, <paragraph> tags but in result files only<head>, <p>can be found. I wonder where the transformation takes place? I’ve checked the saxparser and parser files, but I am still quite confused.
Thanks in advance.
Issue Analytics
- State:
- Created 5 years ago
- Comments:5 (3 by maintainers)
Top Results From Across the Web
Top Related Medium Post
Top Related StackOverflow Question
Troubleshoot Live Code
Top Related Reddit Thread
Top Related Hackernoon Post
Top Related Tweet
Top Related Dev.to Post
Top Related Hashnode Post
Hello !
Training data are in TEI XML (flat TEI preserving stream order of the PDF), there is a mapping between TEI XML tag and labels used by the models in the SAX parsers corresponding to the model in
grobid-trainerThen the final results are serialized into a complex TEI XML (normalised order and more deeply structured). So there is also a mapping between the labels used by the models into this TEI, which is mainly done in the file TEIFormatter, and for substructures like date, person, etc. directly by the POJO classes under
org.grobid.core.data.Hope it makes thing clearer ! If not, don’t hesitate to ask more.
Hi again @Punchwes ! I am closing this issue which was about xml tag/crf label correspondence.
I open a separate one about level and number of section header, to keep track of improvement on this aspect.