Hi, I have 3 different entities and am wondering which config pipeline would be most beneficial for me. Here is an example of the different entities I have:
The object ID will be some random assortment of numbers and letters.
Thanks!
Hi jdugs,
which pipeline you should choose depends mainly on how much training data you have. E.g. if you have few data you should use the spacy pipeline but if you have a lot of data or your data is highly domain-specific you should use the supervised embeddings pipeline. A good way to go about Choosing a Pipeline is to start with one the pre-configured ones and then change components to better suit your needs. Since you are defining custom entities you should use the CRF entity extractor of the Mitie entity extractor (pre-configured pipelines mostly use CRF) Entity Extraction