I poisoned 20 drug names to find out whether a local model belongs in a pharmacy
Описание и применение
Система читает рукописные рецепты, использует локальную модель gemma3:4b для интерпретации названий лекарств, применяет детерминированные проверки безопасности и передаёт только сомнительные строки человеку, тем самым повышая точность и снижая ручной труд в аптеке.
Задача
A neighbourhood medical shop in India needs a reliable way to read handwritten prescriptions and decide which lines can be filled automatically and which require human review, without relying on network connectivity.
Результат
The system processes a prescription in ~58 s on an eight‑core laptop, flags all lines for review due to low confidence, and demonstrates that a local LLM can correctly interpret drug names when combined with deterministic safety checks.
Рабочий процесс
- Photograph handwritten prescription
- Run macOS Vision OCR to extract text
- Send extracted text to local gemma3:4b model via Ollama for drug name interpretation
- Apply deterministic gate checks (safety, grounding, vocabulary)
- Flag lines below confidence threshold or failing checks for human review
- Human reviews flagged lines and completes dispensing
Доказательства
The whole pipeline runs from one command on the shop's machine: python3 scripts/run_page.py prescription.jpg
The OCR reads. The model understands. The gate decides. Only one of those three is allowed to be wrong quietly, and it is not the model.
- Дата записи: 2026-10-17T05:41:25.524Z