Datasets:

Modalities:
Text
Formats:
parquet
DOI:
Libraries:
Datasets
pandas
License:
fdelucaf commited on
Commit
25930de
·
verified ·
1 Parent(s): 3f58e96

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +47 -26
README.md CHANGED
@@ -2,17 +2,15 @@
2
  language:
3
  - ca
4
  - fr
5
- - multilingual
6
  multilinguality:
7
- - translation
8
  pretty_name: CA-FR Parallel Corpus
9
  size_categories:
10
  - 10M<n<100M
11
- source_datasets:
12
- - original
13
  task_categories:
14
  - translation
15
  task_ids: []
 
16
  ---
17
 
18
  # Dataset Card for CA-FR Parallel Corpus
@@ -43,34 +41,47 @@ task_ids: []
43
 
44
  ### Dataset Summary
45
 
46
- The CA-FR Parallel Corpus is a Catalan-French dataset of **18.634.844** 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 French.
56
 
57
  ## Dataset Structure
58
 
59
- Two separated txt files are provided with the sentences sorted in the same order:
 
 
60
 
61
  - ca-fr_corpus.ca: contains 18.634.844 Catalan sentences.
62
 
63
  - ca-fr_corpus.fr: contains 18.634.844 French 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
  |:----------------|----------:|
@@ -87,11 +98,23 @@ The dataset is a combination of the following authentic datasets:
87
 
88
  All corpora were collected from [Opus](https://opus.nlpl.eu/).
89
 
90
- ### Data preparation
 
 
91
 
92
- All datasets are deduplicated and filtered to remove any sentence pairs with a cosine similarity of less than 0.75.
93
- This is done using sentence embeddings calculated using [LaBSE](https://huggingface.co/sentence-transformers/LaBSE).
94
- The filtered datasets are then concatenated to form a final corpus of **18.634.844** parallel sentences.
 
 
 
 
 
 
 
 
 
 
95
 
96
  ### Personal and Sensitive Information
97
 
@@ -101,30 +124,28 @@ No anonymisation process was performed.
101
 
102
  ### Social Impact of Dataset
103
 
104
- The purpose of this dataset is to help develop Machine Translation tasks for mid-resource languages such as Catalan.
105
 
106
  ### Discussion of Biases
107
 
108
- 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.
109
- Nonetheless, we have not applied any steps to reduce their impact.
110
 
111
  ### Other Known Limitations
112
 
113
- 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.
114
 
115
  ## Additional Information
116
 
117
- ### Author
118
- Language Technologies Unit (LangTech) at the Barcelona Supercomputing Center.
 
 
 
119
 
120
- ### Contact information
121
- For further information, please send an email to langtech@bsc.es.
122
 
123
- ### Copyright
124
- Copyright Language Technologies Unit at Barcelona Supercomputing Center (2023).
125
 
126
- ### Licensing information
127
  This work is licensed under a [Attribution-NonCommercial-ShareAlike 4.0 International](https://creativecommons.org/licenses/by-nc-sa/4.0/).
128
 
129
- ### Funding
130
- This work has been promoted and financed by the Generalitat de Catalunya through the [Aina project] (https://projecteaina.cat/).
 
2
  language:
3
  - ca
4
  - fr
 
5
  multilinguality:
6
+ - multilingual
7
  pretty_name: CA-FR Parallel Corpus
8
  size_categories:
9
  - 10M<n<100M
 
 
10
  task_categories:
11
  - translation
12
  task_ids: []
13
+ license: cc-by-nc-sa-4.0
14
  ---
15
 
16
  # Dataset Card for CA-FR Parallel Corpus
 
41
 
42
  ### Dataset Summary
43
 
44
+ The CA-FR Parallel Corpus is a Catalan-French dataset of **18.634.844** parallel sentences.The dataset was created to support Catalan in NLP tasks, specifically
45
  Machine Translation.
46
 
47
  ### Supported Tasks and Leaderboards
48
 
49
+ The dataset can be used to train Bilingual Machine Translation models between <<>> and Catalan in any direction,
50
+ as well as Multilingual Machine Translation models.
51
 
52
  ### Languages
53
 
54
+ The sentences included in the dataset are in Catalan (CA) and French (FR).
55
 
56
  ## Dataset Structure
57
 
58
+ ### Data Instances
59
+
60
+ Two separate txt files are provided with the sentences sorted in the same order:
61
 
62
  - ca-fr_corpus.ca: contains 18.634.844 Catalan sentences.
63
 
64
  - ca-fr_corpus.fr: contains 18.634.844 French sentences.
65
 
66
+ ### Data Fields
67
+
68
+ [N/A]
69
+
70
  ### Data Splits
71
 
72
  The dataset contains a single split: `train`.
73
 
74
  ## Dataset Creation
75
 
76
+ ### Curation Rationale
77
+
78
+ This dataset is aimed at promoting the development of Machine Translation between Catalan and other languages, specifically French.
79
+
80
  ### Source Data
81
 
82
+ #### Initial Data Collection and Normalization
83
+
84
+ The dataset is a combination of the following original datasets:
85
 
86
  | Dataset | Sentences |
87
  |:----------------|----------:|
 
98
 
99
  All corpora were collected from [Opus](https://opus.nlpl.eu/).
100
 
101
+ All datasets are deduplicated and filtered to remove any sentence pairs with a cosine similarity of less than 0.75.
102
+ This is done using sentence embeddings calculated using [LaBSE](https://huggingface.co/sentence-transformers/LaBSE).
103
+ The filtered datasets are then concatenated to form a final corpus of **18.634.844** parallel sentences.
104
 
105
+ #### Who are the source language producers?
106
+
107
+ [Opus](https://opus.nlpl.eu/)
108
+
109
+ ### Annotations
110
+
111
+ #### Annotation process
112
+
113
+ The dataset does not contain any annotations.
114
+
115
+ #### Who are the annotators?
116
+
117
+ [N/A]
118
 
119
  ### Personal and Sensitive Information
120
 
 
124
 
125
  ### Social Impact of Dataset
126
 
127
+ 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.
128
 
129
  ### Discussion of Biases
130
 
131
+ No specific bias mitigation strategies were applied to this dataset.
132
+ Inherent biases may exist within the data.
133
 
134
  ### Other Known Limitations
135
 
136
+ 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.
137
 
138
  ## Additional Information
139
 
140
+ ### Dataset Curators
141
+
142
+ Language Technologies Unit at the Barcelona Supercomputing Center (langtech@bsc.es).
143
+
144
+ 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).
145
 
 
 
146
 
147
+ ### Licensing Information
 
148
 
 
149
  This work is licensed under a [Attribution-NonCommercial-ShareAlike 4.0 International](https://creativecommons.org/licenses/by-nc-sa/4.0/).
150
 
151
+ ### Citation Information