Hey folks! I've written about a spaCy project that uses the SpanRuler and an NER model to find named entities within restaurant reviews. As part of this project, I also reannotated the dataset to fix inconsistencies in the original annotations using the review recipe. This allowed me to compare the original annotations with a trained NER model's annotations to speed up the annotation process and get more consistent results! You can find the full project on the spaCy projects GitHub repo or read about the process on my blog.
3 Likes