TPM-28/Ask_Aegis

Usage example

from transformers import pipeline

classifier = pipeline("text-classification", model="TPM-28/Ask_Aegis")

text = "bonjour j'aurais besoin d'aide pour la config d'un bot."

result = classifier(text)[0]

print(f"Texte: {text}")
print(f"Prédiction: {result['label']}")
print(f"Positive = {result['score']:.2%}, Négative = {1 - result['score']:.2%}")

Metrics

loss: 0.1480395644903183

f1: 0.8282828282828283

precision: 0.82

recall: 0.8367346938775511

auc: 0.9795918367346939

accuracy: 0.9615384615384616