0

According to the Tutorial "Training a custom ENTITY LINKING model with spaCy" (20:33) this is the training data format for spaCy's Entity Linker:

TRAIN_DATA = ("Emerson was born on a farm in Blackbutt, Queensland.", {"links": {(0, 7): { "Q312545": 1.0 }}})

My search for open source annotation tool is not successful. Do you know an open source/free annotation tool for this format?

gzkhv
  • 11
  • 2
  • Do the earlier parts of the video not show you how to setup and annotate data using spaCy's annotation tool Prodigy? – Oxbowerce Oct 14 '21 at 09:41
  • Yes with Prodigy. But it is unfortunately not free or open-source (https://prodi.gy/docs/). – gzkhv Oct 14 '21 at 11:29
  • I would probably use a different open source annotation tool such as [doccano](https://github.com/doccano/doccano) to annotate your data and create a custom script to convert their format to the format spaCy expects. – Oxbowerce Oct 14 '21 at 14:30
  • It would work for me. Thank you! – gzkhv Oct 14 '21 at 15:32

0 Answers0