Excellence in Medical NER Modeling
OpenMed's NER models are designed with a variety of architectures ranging from 65M to 568M parameters, making them particularly suitable for processing complex entities in clinical records and research documents. The SuperClinical series of models can simultaneously identify 17 types of medical entities, including chemical substances, genetic variants, tumor markers, etc., and achieve an accuracy rate of 99.91% when dealing with specialized texts such as "KRAS gene mutation drives tumor formation. The platform has innovatively developed a model discovery application that supports model screening by pharmacology, oncology and other professional fields, which improves the efficiency by more than 3 times compared with the traditional single-model solution. Actual tests show that batch processing of 100 records on the BI55/MedText dataset takes only minutes.
This answer comes from the articleOpenMed: an open source platform for free AI models in healthcareThe
English 
简体中文 
日本語 
Deutsch 
Português do Brasil