everyone-free/hi-bye

hello-or-output

A tiny classifier for a toy chatbot:

You send	It replies
a greeting, e.g. hi or hello how are you	hello
something about computers, e.g. computer or my laptop is slow	output
anything else	I don't understand

How it works

1. all-MiniLM-L6-v2 turns the text into an embedding (a list of numbers that captures its meaning).
2. A logistic regression (classifier.joblib) predicts greeting, computer or other.
3. handler.py maps that label to the reply.

It was trained on a small set of hand-written example sentences. It is a demo, not a production model.

Call the endpoint

curl https://YOUR-ENDPOINT-URL \
  -H "Authorization: Bearer $HF_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"inputs": "hello how are you"}'

Example response:

{"reply": "hello", "label": "greeting", "score": 0.95}