PeterPaker123/Qwen2.5-7B-Vietnamese-Medical-NER-GRPO
PeterPaker123/Qwen2.5-7B-Vietnamese-Medical-NER-GRPO is a 7.6 billion parameter generative language model, based on the Qwen2.5-7B architecture, specifically fine-tuned for Vietnamese Medical Named Entity Recognition (NER). Developed by PeterPaker123, this model utilizes a two-stage Supervised Fine-Tuning (SFT) and Group Relative Policy Optimization (GRPO) pipeline to accurately extract medical entities into a structured JSON format. Its primary use case is dynamic, prompt-defined extraction of medical entities like symptoms, diseases, and drugs from Vietnamese medical texts, offering a generative approach to NER.
Loading preview...
Overview
This model, developed by PeterPaker123, is a specialized generative Large Language Model (LLM) built upon the Qwen2.5-7B architecture. It is meticulously fine-tuned for Vietnamese Medical Named Entity Recognition (NER), moving beyond traditional token-classification methods. The model excels at extracting medical entities such as symptoms, diseases, procedures, and drugs from Vietnamese texts, outputting them in a highly structured JSON format.
Key Capabilities
- Generative NER: Utilizes an LLM as an intelligent extraction agent, processing Vietnamese medical texts and generating structured JSON outputs.
- Dynamic Entity Extraction: Supports prompt-defined entity extraction, allowing users to specify any medical or clinical entity type in the system prompt.
- High Accuracy: Achieves near-perfect Precision, Recall, and Type Accuracy due to a two-stage training pipeline involving Supervised Fine-Tuning (SFT) and Group Relative Policy Optimization (GRPO) with mathematical penalties for hallucinations and missed entities.
- Structured Output: Consistently provides extracted entities in a JSON list format, making it easy for downstream applications.
Good for
- Medical Chatbot Enrichment: Enhancing RAG pipelines for healthcare applications.
- Automated Summarization: Summarizing telehealth transcripts by extracting key medical information.
- Clinical Knowledge Graph Construction: Building structured knowledge bases from unstructured medical text.
- Research and Development: As an AI research tool for advanced Vietnamese medical text processing, with human oversight required for clinical applications.