Me-LLaMA - Medical Foundation Large Language Models for Comprehensive Text Analysis and Beyond

Watch a comprehensive research presentation exploring Me-LLaMA, a suite of medical foundation large language models developed for enhanced medical text analysis. Learn how these models were trained on 129B pre-training tokens and 214K instruction tuning samples from diverse biomedical and clinical sources, requiring over 100,000 A100 GPU hours for the 70B models. Discover how Me-LLaMA outperforms existing open-source medical LLMs in both zero-shot and supervised learning settings across six medical text analysis tasks and 12 benchmark datasets. Understand how task-specific instruction tuning enables Me-LLaMA to surpass commercial LLMs like ChatGPT and GPT-4 on multiple datasets. Presented by Dr. Qianqian Xie, an Associate Research Scientist at Yale University's Department of Biomedical Informatics and Data Science, whose research focuses on natural language processing applications in medicine and has been published in leading conferences and journals.