Prototype
Roman Urdu Medical Triage Research
A private, offline research prototype for classifying Roman Urdu cardiac complaints.
Only the owner can see this repository.
The vision
Study a reproducible language-and-classification pipeline while reporting its limits honestly.
The problem
Roman Urdu complaints are difficult to process consistently in a research workflow.
Our idea
Combine local translation, clinical text embeddings and structured inputs, then compare models on a fixed research dataset.
What it does
- Local research interface and batch processing
- Documented model comparison
- No clinical validation or patient-care use
Interested in this?
Send a short message. We will reply in plain words, no sales pitch.