|
--- |
|
license: mit |
|
task_categories: |
|
- token-classification |
|
tags: |
|
- BioNER |
|
- Biomedical NER |
|
--- |
|
# AIObioEnts model files |
|
|
|
This dataset contains the model files for [AIObioEnts](https://github.com/sirisacademic/AIObioEnts), trained using [AIONER](https://github.com/ncbi/AIONER) with 4 different pre-trained models: |
|
|
|
- [BiomedBERT-base pre-trained on abstracts from PubMed; the best-performing model reported in the original AIONER paper](https://huggingface.co/microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract) |
|
- [BiomedBERT-base pre-trained on both abstracts from PubMed and full-texts articles from PubMedCentral](https://huggingface.co/microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext) |
|
- [BioLinkBERT-base](https://huggingface.co/michiyasunaga/BioLinkBERT-base) |
|
- [BioLinkBERT large](https://huggingface.co/michiyasunaga/BioLinkBERT-large) |
|
|
|
|
|
for the identification of core biomedical entities—gene, disease, cell line, chemical, species, variant— in textual data, in addition to the models fine-tuned with selected entites from the [AnatEM](https://nactem.ac.uk/anatomytagger/#AnatEM) corpus—cell component, tissue, organ, multi-tissue structure, cancer. |
