|
|
--- |
|
|
language: |
|
|
- pl |
|
|
tags: |
|
|
- medical |
|
|
--- |
|
|
PLMedNER is a lightweight Polish-language NER model trained on 50,000 anonymized clinical notes. It recognizes key medical entities such as diagnoses, medications, symptoms, procedures, test results, anatomical terms, clinical recommendations, hospital departments, and patient demographics. |
|
|
|
|
|
🧠 209M parameters – optimized for local/on-premise deployment |
|
|
|
|
|
📄 Supports real-world hospital documents (discharge summaries, consultations, radiology reports) |
|
|
|
|
|
🔐 Fully local processing – privacy-safe and GDPR-compliant |
|
|
|
|
|
⚙️ Easy integration with HIS/EHR systems or research pipelines |
|
|
|
|
|
Entity types: `DISEASE, MEDICATION, SYMPTOM, PROCEDURE, RESULT, ANATOMY, RECOMMENDATION, DEPARTMENT, DEMOGRAPHICS, STAFF` |
|
|
|
|
|
Ideal for clinical NLP tasks in Polish healthcare, including document structuring, research data extraction, and automated anonymization. |
|
|
|
|
|
Read more at https://www.medalion.tech/pl/medner |
