|
--- |
|
license: openrail |
|
language: |
|
- en |
|
pipeline_tag: fill-mask |
|
tags: |
|
- biology |
|
- medical |
|
widget: |
|
- text: "poc all well. wound healed. No [MASK] on exam. Microchip working. Sign off, resee if worried." |
|
example_title: "Post operative Checkup" |
|
- text: "other 2 degu's unwell recently want health check for this one appears well for age blood [MASK] 3.8. offer to reweigh and monitor weight" |
|
example_title: "Blood Glucose check" |
|
--- |
|
|
|
|
|
# VetBERT Pretrained model for Veterinary Clinical Tasks |
|
|
|
This is the pretrained VetBERT model from the github repo: [https://github.com/havocy28/VetBERT](https://github.com/havocy28/VetBERT) |
|
|
|
<!-- Provide a quick summary of what the model is/does. --> |
|
This pretrained model is designed for performing NLP tasks related to veterinary clinical notes. The [Domain Adaptation and Instance Selection for Disease Syndrome Classification over Veterinary Clinical Notes](https://aclanthology.org/2020.bionlp-1.17) (Hur et al., BioNLP 2020) paper introduced VetBERT model: an initialized Bert Model with ClinicalBERT (Bio+Clinical BERT) and further pretrained on the [VetCompass Australia](https://www.vetcompass.com.au/) corpus for performing tasks specific to veterinary medicine. This paper discusses [VetBERTDx](https://huggingface.co/havocy28/VetBERTDx), the finetuned version of VetBERT trained for the the disease classification task. |
|
|
|
## Pretraining Data |
|
|
|
The VetBERT model was initialized from [Bio_ClinicalBERT model](https://huggingface.co/emilyalsentzer/Bio_ClinicalBERT), which was initialized from BERT. The VetBERT model was trained on over 15 million veterinary clincal Records and 1.3 Billion tokens. |
|
|
|
## Pretraining Hyperparameters |
|
|
|
During the pretraining phase for VetBERT, we used a batch size of 32, a maximum sequence length of 512, and a learning rate of 5 · 10−5. The dup factor for duplicating input data with different masks was set to 5. All other default parameters were used (specifically, masked language model probability = 0.15 and max predictions per sequence = 20). |
|
|
|
## VetBERT Finetuning |
|
|
|
VetBERT was further finetuned on a set of 5002 annotated clinical notes to classifiy the disease syndrome associated with the clinical notes as outlined in the paper: [Domain Adaptation and Instance Selection for Disease Syndrome Classification over Veterinary Clinical Notes](https://aclanthology.org/2020.bionlp-1.17) |
|
|
|
## How to use the model |
|
|
|
Load the model via the transformers library: |
|
|
|
``` |
|
from transformers import AutoTokenizer, AutoModelForMaskedLM |
|
|
|
tokenizer = AutoTokenizer.from_pretrained("havocy28/VetBERT") |
|
model = AutoModelForMaskedLM.from_pretrained("havocy28/VetBERT") |
|
|
|
VetBERT_masked = pipeline("fill-mask", model=model, tokenizer=tokenizer) |
|
VetBERT('Suspected pneuomina, will require an [MASK] but in the meantime will prescribed antibiotics') |
|
|
|
``` |
|
|
|
## Citation |
|
|
|
Please cite this article: Brian Hur, Timothy Baldwin, Karin Verspoor, Laura Hardefeldt, and James Gilkerson. 2020. [Domain Adaptation and Instance Selection for Disease Syndrome Classification over Veterinary Clinical Notes](https://aclanthology.org/2020.bionlp-1.17). In Proceedings of the 19th SIGBioMed Workshop on Biomedical Language Processing, pages 156–166, Online. Association for Computational Linguistics. |
|
|
|
|
