Remove case sensitivity when using DIET entity extraction


Can I use the DIET entity extraction and ensure case insensivity? At the moment it recognizes entities when I use caps (as per the training data) but as soon as I don’t use caps etc., it’s poor at extracting. People may use caps or not when they enter their input.

Ideally I want case not to influence this.

Looking at the docs I can’t see a way of ensuring this within the pipeline. Any guidance?


Are you sure it’s DIET that’s doing that?

You can set case_sensitive: False under SpacyNLP, RegexFeaturizer, KeywordIntentClassifier, RegexEntityExtractor.