lisaterumi's picture
ner model
f4ebe87
|
raw
history blame
958 Bytes
---
language: "pt"
widget:
- text: "Dispneia venoso central em subclavia D duplolumen recebendo solução salina e glicosada em BI."
- text: "Paciente com Sepse pulmonar em D8 tazocin (paciente não recebeu por 2 dias Atb)."
- text: "FOI REALIZADO CURSO DE ATB COM LEVOFLOXACINA POR 7 DIAS."
datasets:
- SemClinBr
thumbnail: "https://raw.githubusercontent.com/HAILab-PUCPR/BioBERTpt/master/images/logo-biobertpr1.png"
---
# Portuguese Clinical NER - Quantitative
The Quantitative NER model is part of the [BioBERTpt project](https://www.aclweb.org/anthology/2020.clinicalnlp-1.7/), where 13 models of clinical entities (compatible with UMLS) were trained. All NER model from "pucpr" user was trained from the Brazilian clinical corpus [SemClinBr](https://github.com/HAILab-PUCPR/SemClinBr), with 10 epochs and IOB2 format, from BioBERTpt(all) model.
## Questions?
Post a Github issue on the [BioBERTpt repo](https://github.com/HAILab-PUCPR/BioBERTpt).