johnpaulbin/beanbox-toxic

johnpaulbin/toxic-MiniLM-L6-H384-uncased

Test if a sentence is toxic. Only works for english sentences.

Usage

Basic classification. Labels: [NOT TOXIC, TOXIC]

Install setfit !pip install setfit

from setfit import SetFitModel

model = SetFitModel.from_pretrained("johnpaulbin/beanbox-toxic")

inpt = "" #@param {type:"string"}

out = model.predict_proba([inpt])

if out[0][0] > out[0][1]:
    print("Not toxic")
else:
    print("Toxic!")

print(f"NOT TOXIC: {out[0][0]}\nTOXIC: {out[0][1]}")