alaminxpro/university-complaint-router-llm

AUST Complaint Router — Fine-tuned LLM API

Fine-tuned Qwen3.5-0.8B (Q4KM quantization) running on CPU via llama-cpp-python.

Model

API

GET /health

Health check — also wakes the Space from sleep.

POST /university-complain-router/

Classify a complaint.

curl -X POST https://alaminxpro-aust-complaint-router-llm.hf.space/university-complain-router/ \
  -H "Content-Type: application/json" \
  -d '{"complaint": "The projector in room 4A05 is broken."}'

Swagger docs: /docs

Note

This is the LLM-based endpoint (~5-15s per request on free CPU). For faster inference (<2ms), see the TF-IDF + SVM version.