Datasets:

Modalities:
Text
Formats:
parquet
DOI:
Libraries:
Datasets
pandas
License:
fdelucaf commited on
Commit
7589c1b
·
verified ·
1 Parent(s): 3f6711b

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +54 -27
README.md CHANGED
@@ -2,17 +2,15 @@
2
  language:
3
  - ca
4
  - it
5
- - multilingual
6
  multilinguality:
7
- - translation
8
  pretty_name: CA-IT Parallel Corpus
9
  size_categories:
10
  - 1M<n<10M
11
- source_datasets:
12
- - original
13
  task_categories:
14
  - translation
15
  task_ids: []
 
16
  ---
17
 
18
  # Dataset Card for CA-IT Parallel Corpus
@@ -41,36 +39,51 @@ task_ids: []
41
 
42
  ## Dataset Description
43
 
 
 
44
  ### Dataset Summary
45
 
46
- The CA-IT Parallel Corpus is a Catalan-Italian dataset of **9.482.927** parallel sentences. The dataset was created to support Catalan NLP tasks, e.g.,
47
  Machine Translation.
48
 
49
  ### Supported Tasks and Leaderboards
50
 
51
- The dataset can be used to train a model for Multilingual Machine Translation. Success on this task is typically measured by achieving a high BLEU score.
 
52
 
53
  ### Languages
54
 
55
- The texts in the dataset are in Catalan and Italian.
56
 
57
  ## Dataset Structure
58
 
59
- Two separated txt files are provided with the sentences sorted in the same order:
 
 
60
 
61
  - ca-it_corpus.ca: contains 9.482.927 Catalan sentences.
62
 
63
  - ca-it_corpus.it: contains 9.482.927 Italian sentences.
64
 
 
 
 
 
65
  ### Data Splits
66
 
67
  The dataset contains a single split: `train`.
68
 
69
  ## Dataset Creation
70
 
 
 
 
 
71
  ### Source Data
72
 
73
- The dataset is a combination of the following authentic datasets:
 
 
74
 
75
  | Dataset | Sentences |
76
  |:--- | ---: |
@@ -87,45 +100,59 @@ The dataset is a combination of the following authentic datasets:
87
 
88
  All corpora were collected from [Opus](https://opus.nlpl.eu/).
89
 
 
 
 
90
 
91
- ### Data preparation
92
 
93
- All datasets are deduplicated and filtered to remove any sentence pairs with a cosine similarity of less than 0.75.
94
- This is done using sentence embeddings calculated using [LaBSE](https://huggingface.co/sentence-transformers/LaBSE).
95
- The filtered datasets are then concatenated to form a final corpus of **9.482.927** parallel sentences.
96
 
97
- ### Personal and Sensitive Information
98
 
 
 
 
 
 
 
 
 
 
99
  No anonymisation process was performed.
100
 
101
  ## Considerations for Using the Data
102
 
103
  ### Social Impact of Dataset
104
 
105
- The purpose of this dataset is to help develop Machine Translation tasks for mid-resource languages such as Catalan.
106
 
107
  ### Discussion of Biases
108
 
109
- We are aware that since part of the data comes from unreliable web pages and non-curated texts, some biases may be present in the dataset.
110
- Nonetheless, we have not applied any steps to reduce their impact.
111
 
112
  ### Other Known Limitations
113
 
114
- The dataset contains data of a general domain. Application of this dataset in more specific domains such as biomedical, legal etc. would be of limited use.
115
 
116
  ## Additional Information
117
 
118
- ### Author
119
- Language Technologies Unit (LangTech) at the Barcelona Supercomputing Center.
120
 
121
- ### Contact information
122
- For further information, please send an email to langtech@bsc.es.
123
 
124
- ### Copyright
125
- Copyright Language Technologies Unit at Barcelona Supercomputing Center (2023).
 
 
126
 
127
- ### Licensing information
128
  This work is licensed under a [Attribution-NonCommercial-ShareAlike 4.0 International](https://creativecommons.org/licenses/by-nc-sa/4.0/).
129
 
130
- ### Funding
131
- This work has been promoted and financed by the Generalitat de Catalunya through the [Aina project] (https://projecteaina.cat/).
 
 
 
 
 
 
2
  language:
3
  - ca
4
  - it
 
5
  multilinguality:
6
+ - multilingual
7
  pretty_name: CA-IT Parallel Corpus
8
  size_categories:
9
  - 1M<n<10M
 
 
10
  task_categories:
11
  - translation
12
  task_ids: []
13
+ license: cc-by-nc-sa-4.0
14
  ---
15
 
16
  # Dataset Card for CA-IT Parallel Corpus
 
39
 
40
  ## Dataset Description
41
 
42
+ - **Point of Contact:** langtech@bsc.es
43
+
44
  ### Dataset Summary
45
 
46
+ The CA-IT Parallel Corpus is a Catalan-Italian dataset of **9.482.927** parallel sentences. The dataset was created to support Catalan in NLP tasks, specifically
47
  Machine Translation.
48
 
49
  ### Supported Tasks and Leaderboards
50
 
51
+ The dataset can be used to train Bilingual Machine Translation models between <<>> and Catalan in any direction,
52
+ as well as Multilingual Machine Translation models.
53
 
54
  ### Languages
55
 
56
+ The sentences included in the dataset are in Catalan (CA) and Italian (IT).
57
 
58
  ## Dataset Structure
59
 
60
+ ### Data Instances
61
+
62
+ Two separate txt files are provided with the sentences sorted in the same order:
63
 
64
  - ca-it_corpus.ca: contains 9.482.927 Catalan sentences.
65
 
66
  - ca-it_corpus.it: contains 9.482.927 Italian sentences.
67
 
68
+ ### Data Fields
69
+
70
+ [N/A]
71
+
72
  ### Data Splits
73
 
74
  The dataset contains a single split: `train`.
75
 
76
  ## Dataset Creation
77
 
78
+ ### Curation Rationale
79
+
80
+ This dataset is aimed at promoting the development of Machine Translation between Catalan and other languages, specifically Italian.
81
+
82
  ### Source Data
83
 
84
+ #### Initial Data Collection and Normalization
85
+
86
+ The dataset is a combination of the following original datasets:
87
 
88
  | Dataset | Sentences |
89
  |:--- | ---: |
 
100
 
101
  All corpora were collected from [Opus](https://opus.nlpl.eu/).
102
 
103
+ All datasets are deduplicated and filtered to remove any sentence pairs with a cosine similarity of less than 0.75.
104
+ This is done using sentence embeddings calculated using [LaBSE](https://huggingface.co/sentence-transformers/LaBSE).
105
+ The filtered datasets are then concatenated to form a final corpus of **9.482.927** parallel sentences.
106
 
107
+ #### Who are the source language producers?
108
 
109
+ [Opus](https://opus.nlpl.eu/)
 
 
110
 
111
+ ### Annotations
112
 
113
+ #### Annotation process
114
+
115
+ The dataset does not contain any annotations.
116
+
117
+ #### Who are the annotators?
118
+
119
+ [N/A]
120
+
121
+ ### Personal and Sensitive Information
122
  No anonymisation process was performed.
123
 
124
  ## Considerations for Using the Data
125
 
126
  ### Social Impact of Dataset
127
 
128
+ By providing this resource, we intend to promote the use of Catalan across NLP tasks, thereby improving the accessibility and visibility of the Catalan language.
129
 
130
  ### Discussion of Biases
131
 
132
+ No specific bias mitigation strategies were applied to this dataset.
133
+ Inherent biases may exist within the data.
134
 
135
  ### Other Known Limitations
136
 
137
+ The dataset contains data of a general domain. Applications of this dataset in more specific domains such as biomedical, legal etc. would be of limited use.
138
 
139
  ## Additional Information
140
 
141
+ ### Dataset Curators
 
142
 
143
+ Language Technologies Unit at the Barcelona Supercomputing Center (langtech@bsc.es).
 
144
 
145
+ This work was funded by the [Departament de la Vicepresidència i de Polítiques Digitals i Territori de la Generalitat de Catalunya](https://politiquesdigitals.gencat.cat/ca/inici/index.html#googtrans(ca|en) within the framework of [Projecte AINA](https://politiquesdigitals.gencat.cat/ca/economia/catalonia-ai/aina).
146
+
147
+
148
+ ### Licensing Information
149
 
 
150
  This work is licensed under a [Attribution-NonCommercial-ShareAlike 4.0 International](https://creativecommons.org/licenses/by-nc-sa/4.0/).
151
 
152
+ ### Citation Information
153
+
154
+ [N/A]
155
+
156
+ ### Contributions
157
+
158
+ [N/A]