NER (named entity recognition) is a natural language processing technique that aims to identify and classify entities such as names of people, places, organizations, etc. in unstructured text.
Well, our algorithms and data department has developed its own version!
Loïck BOURDOIS, Data Scientist NLP at CATIE, worked on this type of model by:
– Training models on more data and without leakage (the previous state of the art was trained on leaked data);
– Making the training dataset, FrenchNER, available to French speakers;
– Offering two configurations, base and large, for greater accuracy.
This tool can be used to analyze emails, quickly find information in dense text or anonymize data.
In a few seconds, you can spot places (LOC), people (PER) and organizations (ORG) in your text. Handy!
If you are interested, you can try the tool right now on our Hugging Face account: https://lnkd.in/gpee7MZS