|
|
|
# PubMedBERT Abstract + Full Text Fine-Tuned on QNLI Task |
|
|
|
Use case: You can use it to search through a document for a given question, to see if your question is answered in that document. |
|
|
|
LABEL0 is "not entailment" meaning your question is not answered by the context and LABEL1 is "entailment" meaning your question is answered. |
|
|
|
> Example input: [CLS] Your question [SEP] The context to be searched in [SEP] |
|
|
|
|
|
Link to the original model: https://huggingface.co/microsoft/BiomedNLP-PubMedBERT-base-uncased-abstract-fulltext |
|
|
|
|
|
Credits to the paper: |
|
|
|
> @misc{pubmedbert, author = {Yu Gu and Robert Tinn and Hao Cheng and |
|
> Michael Lucas and Naoto Usuyama and Xiaodong Liu and Tristan Naumann |
|
> and Jianfeng Gao and Hoifung Poon}, title = {Domain-Specific |
|
> Language Model Pretraining for Biomedical Natural Language |
|
> Processing}, year = {2020}, eprint = {arXiv:2007.15779}, } |
|
|