metadata
dataset_info:
features:
- name: id
dtype: string
- name: source
dtype: string
- name: target
dtype: string
splits:
- name: train
num_bytes: 4259301
num_examples: 1563
download_size: 2181118
dataset_size: 4259301
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
task_categories:
- zero-shot-classification
language:
- uk
pretty_name: UA-GEC
size_categories:
- 1K<n<10K
Dataset Card for UA-GEC
This is a revised version of the document-level Ukrainian Grammatical Error Correction (UA-GEC) dataset intended for internal use in the HPLT project. The dataset is constructed as follows:
- The examples are extracted using the library. The source document can appear more than once.
- Examples where the source is the same as the target are removed.
- The remaining training and test examples are combined into one split.
- Note: no clear evaluation script is available, and the target metric is tricky to work with (ERRANT). Hence, the task is formulated as document ranking (the model is required to prefer the grammatical text version to ungrammatical one).
Dataset Details
Dataset Sources
- Repository: github.com/grammarly/ua-gec
- Paper: arxiv.org/abs/2103.16997
Dataset Instance
{
'id': '0183',
'source': 'Останнім часом вони мене стали дратувати. Літають там собі на своїх літаючих тарілках і горя не знають, а у мене програма не працює. Прилетіли б от, та й відлагодили її за мене. Ну, самі, або своїх роботів підписали на справу - це мені без різниці. Я б в загальних рисах пояснив, як воно повинно працювати, а вони б все підправили.\nОсь! А я б в цей час на диванчику повалявся.',
'target': 'Останнім часом вони мене стали дратувати. Літають там собі на своїх тарілках і горя не знають, а у мене програма не працює. Прилетіли б от та й відлагодили її за мене. Ну, самі, або своїх роботів підписали на справу — це мені без різниці. Я б у загальних рисах пояснив, як воно повинно працювати, а вони б усе підправили.\nОсь! А я б у цей час на диванчику повалявся.'
}
Dataset Structure
id
: the example id; source
: the source document (the ungrammatical text version); target
: the target document (the grammatical text version).