|
--- |
|
license: creativeml-openrail-m |
|
--- |
|
metadata |
|
tags: |
|
- T5 |
|
- transformers |
|
- Question Answering |
|
- multilingual |
|
- en |
|
- ru |
|
- uk |
|
- pl |
|
|
|
In [this article](https://medium.com/@uaritm/why-is-there-so-much-noise-around-one-of-the-neural-network-models-5d58cac8d706), you will find information about applying this model in the [clinic search application](https://aihealth.site/) |
|
|
|
|
|
uaritm/T5_ukruen_qa_all_clean_10 |
|
|
|
The model is trained on a question-answer dataset (250000 questions to doctors of various specialties and short answers from doctors). |
|
Texts in four languages (English, Russian, Ukrainian and Polish). |
|
|
|
Model trained 10 epochs based on the model 'uaritm/lik_neuro_202' |
|
|
|
You can talk about your health problems and this neural network will give you advice. |
|
You can see how the model works and test it at the link: https://aihealth.site |
|
More detailed information about this neural network can be found here: https://www.esemi.org/vgp-the-new-online-resource-for-medical-assistance/ |
|
|
|
|
|
|
|
from transformers import T5ForConditionalGeneration, T5Tokenizer |
|
import torch |
|
|
|
|
|
Citing & Authors |
|
@misc{Uaritm, |
|
title={SetFit: Question Answering with medical context}, |
|
author={Vitaliy Ostashko}, |
|
year={2023}, |
|
url={https://aihealth.site}, |
|
} |
