EPFLiGHT/Meditron3-Gemma2-9B

TEXT GENERATIONConcurrency Cost:1Model Size:9BQuant:FP8Ctx Length:16kPublished:Jan 16, 2025License:gemmaArchitecture:Transformer0.0K Cold

EPFLiGHT/Meditron3-Gemma2-9B is a 9 billion parameter causal decoder-only transformer language model developed by the OpenMeditron initiative, specialized in clinical medicine. Built upon the Gemma2-9B base model, it emphasizes equitable representation, contextual diversity, and real-world evidence-based guidelines, particularly for limited-resource and humanitarian settings. This model is designed for research into enhancing clinical decision-making and access to medical information, with a context length of 16384 tokens.

Loading preview...

Meditron3-Gemma2-9B: A Medical LLM for Research

EPFLiGHT/Meditron3-Gemma2-9B is a 9 billion parameter large language model developed by the OpenMeditron initiative, specifically designed for clinical medicine. It is built on the Gemma2-9B base model and focuses on equitable representation, contextual diversity, and actionable evidence-based guidelines, with a particular emphasis on limited-resource and humanitarian settings.

Key Capabilities & Specialization

  • Medical Specialization: Trained extensively on a diverse medical dataset including clinical guidelines, peer-reviewed publications, synthetic differential diagnoses, and LLM-enhanced medical MCQs.
  • Research Focus: Primarily intended as a research tool to explore the potential of LLMs in clinical decision-making and medical information access.
  • Foundation Model: Provided as a foundation model, it is not instruction-tuned but can be adapted for downstream tasks like question-answering through in-context learning or fine-tuning.
  • Performance: Shows modest improvements over the base Gemma2-9B model on medical multiple-choice question benchmarks like MedmcQA, MedQA, PubmedQA, and notably on AfrimedQA.

Intended Use Cases

  • Medical Research: Ideal for academic and research purposes to study LLM applications in medicine.
  • Downstream Adaptation: Suitable for developers looking to fine-tune a medical foundation model for specific clinical or healthcare-related applications.

Important Note: This model is research-only and not validated for direct medical use or clinical decision-making. It is provided "as is" under the Gemma Model License Agreement.