MobileBERT for Phishing Email Detection

This model is a fine-tuned MobileBERT for binary phishing-vs-legitimate email classification.

Usage

Python (transformers):

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

tok = AutoTokenizer.from_pretrained("Perth0603/phishing-email-mobilebert")
mdl = AutoModelForSequenceClassification.from_pretrained("Perth0603/phishing-email-mobilebert")
inputs = tok(["Win an iPhone! Click here"], return_tensors="pt")
with torch.no_grad():
    pred = mdl(**inputs).logits.softmax(-1)
print(pred)

HF Inference API (public models only):

curl -X POST https://api-inference.huggingface.co/models/Perth0603/phishing-email-mobilebert   -H "Content-Type: application/json"   -d '{"inputs": "Win an iPhone! Click here"}'