1aurent

/

distilbert-base-multilingual-cased-finetuned-email-spam

like 3

I'm trying to classify spam using this model. I'm passing the content like this:
Title: XXX
Content: XXX
But in practice, it identifies an unjustifiably high percentage of content as spam.
Are there any specific guidelines on how to use the model? Does the model expect input in a specific format?

Thank you.