top of page

🐝⛑️It’s here…meet Meditron-3!

Check out our model release HERE






Meditron is the latest update to our suite of state-of-the-art open Large Medical Language Models trained from Llama-3.1 [8B] and [70B]


---


💯 OPEN weights released here 🔗 https://lnkd.in/eF8q9A94


🤝CO-DESIGNED: from day one with an interdisciplinary team of practicing clinicians from around the world, global humanitarian organizations, AI ethicists, and data scientists


📊STATE OF THE ART: Out-performing GPT4 on the standard medical benchmarks (*see footnote)


🌍REPRESENTATIVE: Continually pre-trained on an enormous corpus of expert-curated open-access medical literature and global clinical practice guidelines that is specifically inclusive of low-resource and humanitarian settings


🤖🤺 join a diverse global community of clinical experts 👩🏿‍⚕️🧑🏻‍⚕️👨🏼‍⚕️👩🏽‍⚕️🧑🏾‍⚕️👩🏼‍⚕️👨🏻‍⚕️ in the MOOVE. Sign up to show your interest here 🔗 https://lnkd.in/eQb3KyDc


--


Built with ❤️ by an incredible and mostly volunteer collaborative of 

--💼Students and researchers at EPFLYale School of Medicine, and Carnegie Mellon University Africa

--⛑️Humanitarian advisors at the International Committee of the Red Cross - ICRC 

--🏥Clinicians from across the world from Nairobi to Lausanne!

🏷️ Tagged in the comments by our phenomenal team lead Alexandre Sallinen 

I am extremely lucky and grateful to work with you all ❤️ 



(*) While these benchmarks are helpful indicators of performance, we certainly can’t rely on results from static, hackable, generic multiple-choice questions from medical exams with a skewed geographic representation. medicineisnotmultiplechoice


We need robust, transparent real world representative expert evaluation in massive-scale clinical trials i.e.#jointhemoove 🇷🇼🇹🇿🇨🇭🇰🇪🇳🇬🇿🇦🇺🇸🇺🇬🇨🇦…Greetings from Kigali!



More models coming soon!

Σχόλια


bottom of page