adityak74/medfit-llm-3B
Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:3.2BQuant:BF16Ctx Length:32kPublished:Sep 15, 2025License:llama3.2Architecture:Transformer0.0K Warm

MEDFIT-LLM-3B is a 3.2 billion parameter causal language model fine-tuned from Meta's Llama-3.2-3B-Instruct by Aditya Karnam Gururaj Rao, Arjun Jaggi, and Sonam Naidu. Optimized for healthcare and medical question-answering, this model demonstrates a 30 percentage point improvement in direct answer capabilities. It excels at providing structured medical information for patient education and healthcare chatbot applications, leveraging a 32768 token context length.

Loading preview...

Overview

MEDFIT-LLM-3B is a specialized 3.2 billion parameter language model, fine-tuned from Meta's Llama-3.2-3B-Instruct. Developed by Aditya Karnam Gururaj Rao, Arjun Jaggi, and Sonam Naidu, this model is specifically optimized for healthcare and medical question-answering (QA) applications. It leverages LoRA (Low-Rank Adaptation) techniques on a curated dataset of 6,444 healthcare-related question-answer pairs, enhancing its ability to provide direct and structured medical information.

Key Capabilities & Performance

  • Enhanced Direct Answers: Achieves a 30 percentage point increase in direct answer rate (from 6.0% to 36.0%) compared to the base model.
  • Improved Response Structure: Shows an 18% increase in the use of numbered lists, leading to better organized and more comprehensive answers.
  • Domain Specialization: Optimized for medical information dissemination and patient education within healthcare chatbot contexts.
  • Efficient Fine-tuning: Utilizes LoRA with the MLX framework, minimizing computational requirements while maximizing domain-specific performance.

Ideal Use Cases

MEDFIT-LLM-3B is designed for applications requiring accurate and structured medical information delivery. It is particularly well-suited for:

  • Medical Question Answering: Providing precise responses to healthcare queries.
  • Patient Education: Delivering easy-to-understand medical information.
  • Healthcare Chatbot Applications: Serving as the core for conversational agents in healthcare.
  • Integration into: Healthcare mobile apps, medical information systems, and telemedicine platforms.

Important Limitations

It is crucial to note that MEDFIT-LLM-3B is not intended for medical diagnosis, treatment recommendations, or emergency situations. It should always be used as a supplementary tool with human oversight and never as a substitute for professional medical advice.