|
--- |
|
language: |
|
- uk |
|
tags: |
|
- text2text-generation |
|
- flair |
|
library_name: generic |
|
license: mit |
|
metrics: |
|
- perplexity |
|
datasets: |
|
- ubertext2.0 |
|
widget: |
|
- text: "Росія зазнає поразки" |
|
- text: "Достеменно відомо, що Україна перемагає" |
|
--- |
|
|
|
# Ukrainian flair embeddings (forward) |
|
|
|
Trained for 10 epochs on the texts from ubertext2.0 (WIP). |
|
The characters dictionary used for training is in `flair_dictionary.pkl` file |
|
|
|
For more information on flair embeddings see [the article](https://github.com/flairNLP/flair/blob/master/resources/docs/embeddings/FLAIR_EMBEDDINGS.md) or the paper below: |
|
|
|
|
|
```bibtex |
|
@inproceedings{akbik2018coling, |
|
title={Contextual String Embeddings for Sequence Labeling}, |
|
author={Akbik, Alan and Blythe, Duncan and Vollgraf, Roland}, |
|
booktitle = {{COLING} 2018, 27th International Conference on Computational Linguistics}, |
|
pages = {1638--1649}, |
|
year = {2018} |
|
} |
|
``` |
|
|
|
Copyright: Dmytro Chaplynskyi, [lang-uk](https://lang.org.ua) project, 2022 |
