From NobodyWho · Duarte O.Carmo · · 50s
Jev in 25 lines of Python
A 25-line Python model that classifies emails as legitimate, spam, or phishing without sending data.
In brief
A 25-line Python model that classifies emails as legitimate, spam, or phishing without sending data. Jev uses local Python code to classify emails with probabilities without external data sharing. Originally reported by NobodyWho.
Jev's Python code
Jev is a 25-line Python model that classifies emails as legitimate, spam, or phishing without sending data.
Load the model
Jev loads a Qwen3-0.6B-GGUF model from Hugging Face using llama-cpp-python.
Define choices
Jev uses labels and choices to prompt email classification with legitimate, spam, or phishing options.
Process email
Jev evaluates the email prompt to output probabilities for legitimate, spam, or phishing emails.
Output probabilities
Jev shows probabilities: legitimate 0.031, spam 0.084, phishing 0.885.
What Jev is
Jev is a local, fast model that classifies emails without sending data to external servers.
What matters most
Jev uses local Python code to classify emails with probabilities without external data sharing.
