microsoft/MediPhi-Instruct

microsoft/MediPhi-Instruct is a 3.8 billion parameter Phi3-based small language model developed by Microsoft Healthcare & Life Sciences, specifically fine-tuned for medical and clinical natural language processing tasks. It was created using a modular approach, merging five expert models trained on diverse medical corpora, and then clinically aligned with the large-scale MediFlow dataset. This model is optimized for research in medically adapted language models, particularly in memory/compute constrained and latency-bound clinical environments.