Datasets:

Modalities:
Text
Formats:
parquet
ArXiv:
Libraries:
Datasets
pandas
License:
murodbek commited on
Commit
9804de3
1 Parent(s): 45b4c56

adding README file

Browse files
Files changed (1) hide show
  1. README.md +115 -0
README.md CHANGED
@@ -30,4 +30,119 @@ configs:
30
  path: data/kaa_rus-*
31
  - split: kaa_uzb
32
  path: data/kaa_uzb-*
 
 
 
 
 
 
 
 
 
 
 
 
 
 
33
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
30
  path: data/kaa_rus-*
31
  - split: kaa_uzb
32
  path: data/kaa_uzb-*
33
+ language:
34
+ - en
35
+ - ru
36
+ - uz
37
+ - kaa
38
+ pretty_name: dilmash
39
+ size_categories:
40
+ - 100K<n<1M
41
+ license: mit
42
+ task_categories:
43
+ - translation
44
+ tags:
45
+ - dilmash
46
+ - karakalpak
47
  ---
48
+ # Dilmash: Karakalpak Parallel Corpus
49
+
50
+ This repository contains a parallel corpus for the Karakalpak language, developed as part of the research paper "Open Language Data Initiative: Advancing Low-Resource Machine Translation for Karakalpak".
51
+
52
+ ## Dataset Description
53
+
54
+ The Karakalpak Parallel Corpus is a collection of 300,000 sentence pairs, designed to support machine translation tasks involving the Karakalpak language. It includes:
55
+
56
+ - Uzbek-Karakalpak (100,000 pairs)
57
+ - Russian-Karakalpak (100,000 pairs)
58
+ - English-Karakalpak (100,000 pairs)
59
+
60
+ ## Usage
61
+
62
+ This dataset is intended for training and evaluating machine translation models involving the Karakalpak language.
63
+
64
+ To load and use dataset, run this script:
65
+
66
+ ```python
67
+ from datasets import load_dataset
68
+
69
+ dilmash_corpus = load_dataset("tahrirchi/dilmash")
70
+ ```
71
+
72
+ ## Dataset Structure
73
+
74
+ ### Data Instances
75
+
76
+ - **Size of downloaded dataset files:** 77.4 MB
77
+ - **Size of the generated dataset:** 46.1 MB
78
+ - **Total amount of disk used:** 123.5 MB
79
+
80
+ An example of 'kaa_eng' looks as follows.
81
+ ```
82
+ {'src_lang': 'kaa_Latn',
83
+ 'src_sent': 'Pedagogikalıq ideal balaǵa ıktıyatlılıq penen katnasta bolıw principine bárqulla, úlken hám kishi jumıslarda súyeniwdi talan etedi.',
84
+ 'tgt_lang': 'eng_Latn',
85
+ 'tgt_sent': 'The ideal of education demands that the principle of treating children with care be observed at all times, in both big and small matters.'
86
+ }
87
+ ```
88
+
89
+ ### Data Fields
90
+
91
+ The data fields are the same among all splits.
92
+
93
+ - `src_lang`: a `string` feature that contains source language.
94
+ - `src_sent`: a `string` feature that contains sentence in source language.
95
+ - `tgt_lang`: a `string` feature that contains target language.
96
+ - `tgt_sent`: a `string` feature that contains sentence in target language.
97
+
98
+ ### Data Splits
99
+
100
+ | split_name |num_examples|
101
+ |-----------------|-----------:|
102
+ | kaa_eng | 100000 |
103
+ | kaa_rus | 100000 |
104
+ | kaa_uzb | 100000 |
105
+
106
+ ## Data Sources
107
+
108
+ The corpus comprises diverse parallel texts sourced from multiple domains:
109
+
110
+ - 23% sentences from news sources
111
+ - 34% sentences from books (novels, non-fiction)
112
+ - 24% sentences from bilingual dictionaries
113
+ - 19% sentences from school textbooks
114
+
115
+ Additionally, 4,000 English-Karakalpak pairs were sourced from the Gatitos Project (Jones et al., 2023)[https://aclanthology.org/2023.emnlp-main.26].
116
+
117
+ ## Data Preparation
118
+
119
+ The data mining process involved local mining techniques, ensuring that parallel sentences were extracted from translations of the same book, document, or article. Sentence alignment was performed using LaBSE (Language-agnostic BERT Sentence Embedding) embeddings.
120
+
121
+ ## Citation
122
+
123
+ If you use this dataset in your research, please cite our paper:
124
+
125
+ ```bibtex
126
+ @inproceedings{mamasaidov2024advancing,
127
+ title={Open Language Data Initiative: Advancing Low-Resource Machine Translation for Karakalpak},
128
+ author={Mamasaidov, Mukhammadsaid and Shopulatov, Abror},
129
+ booktitle={Proceedings of the OLDI Workshop},
130
+ year={2024}
131
+ }
132
+ ```
133
+
134
+ ## Gratitude
135
+
136
+ We are thankful to these awesome organizations and people for helping to make it happen:
137
+
138
+ - [David Dalé](https://daviddale.ru): for advise throughout the process
139
+ - Perizad Najimova: for expertise and assistance with the Karakalpak language
140
+ - [Nurlan Pirjanov](https://www.linkedin.com/in/nurlan-pirjanov/): for expertise and assistance with the Karakalpak language
141
+ - [Atabek Murtazaev](https://www.linkedin.com/in/atabek/): for advise throughout the process
142
+ - Ajiniyaz Nurniyazov: for advise throughout the process
143
+
144
+ ## Contacts
145
+
146
+ We believe that this work will enable and inspire all enthusiasts around the world to open the hidden beauty of low-resource languages, in particular Karakalpak.
147
+
148
+ For further development and issues about the dataset, please use m.mamasaidov@tahrirchi.uz or a.shopolatov@tahrirchi.uz to contact.