mrm8488

/

bert-tiny-finetuned-sms-spam-detection

like 46

It'll return two labels with scores each defining the probability that it is either spam or not spam.

As I understand it, the first value in these spam-detectors is usually the "not spam" probability and the second one the "spam" probability. By testing it with some entries from the training set, I'm pretty sure that this is true for this model as well, so LABEL_0 is "not spam" probabiliyt and LABEL_1 is "spam" probability.

Gotcha! Thank you for your reply!