OnDeviceMedNotes/Medical_Summary_Notes
OnDeviceMedNotes/Medical_Summary_Notes is a 1 billion parameter instruction-tuned causal language model, fine-tuned from meta-llama/Llama-3.2-1B-Instruct using PEFT and LoRA. This model is specifically optimized for text generation, excelling at converting free-text medical transcripts into structured medical notes. It generates comprehensive SOAP notes with predefined sections like Demographics, Presenting Illness, History, and Physical Exam Findings, making it ideal for clinical documentation automation.
Loading preview...
OnDeviceMedNotes/Medical_Summary_Notes: Structured Medical Note Generation
This model, built upon the meta-llama/Llama-3.2-1B-Instruct base, is a 1 billion parameter language model specifically fine-tuned for generating structured medical notes. Utilizing PEFT (Parameter-Efficient Fine-Tuning) with LoRA and the Unsloth library for accelerated training, it transforms free-text medical transcripts into organized clinical documentation.
Key Capabilities
- Input: Processes free-text medical transcripts, such as doctor-patient conversations or dictated notes.
- Output: Generates structured medical notes following a comprehensive SOAP (Subjective, Objective, Assessment, Plan) format.
- Structured Sections: Produces notes with clearly defined sections including Demographics, Presenting Illness, History of Presenting Illness, Past Medical History, Surgical History, Family History, Social History, Allergy History, Medication History, Dietary History, Review of Systems, Physical Exam Findings, Labs and Imaging, Assessment, and Plan.
- Clinical Relevance: Designed to extract and organize critical information from unstructured medical dialogue, enhancing efficiency in clinical settings.
Intended Use Cases
- Automated Clinical Documentation: Streamlines the creation of detailed patient records from raw medical conversations.
- Medical Transcription Processing: Converts spoken or dictated medical notes into a standardized, readable format.
- Healthcare AI Applications: Serves as a foundational component for systems requiring structured medical data extraction and summarization.
For more technical details, refer to the associated research papers: https://arxiv.org/abs/2507.03033 and https://www.medrxiv.org/content/10.1101/2025.07.01.25330679v1.