|
|
--- |
|
|
pipeline_tag: translation |
|
|
library_name: transformers |
|
|
--- |
|
|
|
|
|
### Biomedical French to English Neural Machine Translation |
|
|
|
|
|
<u>Source language:</u> fr |
|
|
<u>Target language:</u> en |
|
|
<u>Training dataset:</u> WMT20, Cochrane bilingual parallel corpus, Taus Corona Crisis corpus, Mlia Covid corpus |
|
|
<u>Development set:</u> Medline 18, Medline 19 |
|
|
<u>Test set:</u> Medline 20 |
|
|
<u>Model:</u> transformer |
|
|
<u>Pre-processing:</u> SentencePiece |
|
|
|
|
|
## Benchmark |
|
|
|
|
|
<div style="width: 20%; text-align: left;"> |
|
|
|
|
|
| **Test set** | **BLEU** | |
|
|
|----------------|----------| |
|
|
| Medline20 | 35.8 | |
|
|
|
|
|
</div> |
|
|
|
|
|
## How to use this Model? |
|
|
* This model can be accessed via git clone: |
|
|
``` |
|
|
git clone https://huggingface.co/SLPG/Biomedical_French_to_English |
|
|
``` |
|
|
* You can use Fairseq library to access the model for translations: |
|
|
``` |
|
|
from fairseq.models.transformer import TransformerModel |
|
|
``` |
|
|
* Load the model |
|
|
``` |
|
|
model = TransformerModel.from_pretrained('path/to/model') |
|
|
``` |
|
|
* Set the model to evaluation mode |
|
|
``` |
|
|
model.eval() |
|
|
``` |
|
|
* Perform inference |
|
|
``` |
|
|
input_text = 'Hello, how are you?' |
|
|
output_text = model.translate(input_text) |
|
|
print(output_text) |
|
|
``` |
|
|
|
|
|
## Citation |
|
|
|
|
|
**If you use our model, kindly cite our [paper](https://hal.science/hal-03430610/document)**: |
|
|
``` |
|
|
@inproceedings{xu2021lisn, |
|
|
title={LISN@ WMT 2021}, |
|
|
author={Xu, Jitao and Rauf, Sadaf Abdul and Pham, Minh Quang and Yvon, Fran{\c{c}}ois}, |
|
|
booktitle={6th Conference on Statistical Machine Translation}, |
|
|
year={2021} |
|
|
} |
|
|
``` |
