Datasets:

Modalities:
Text
Formats:
parquet
DOI:
Libraries:
Datasets
pandas
License:
fdelucaf commited on
Commit
ce4fc37
·
verified ·
1 Parent(s): f90136c

Updated description of the dataset format and content

Browse files
Files changed (1) hide show
  1. README.md +12 -24
README.md CHANGED
@@ -22,7 +22,7 @@ license: cc-by-nc-sa-4.0
22
 
23
  ### Dataset Summary
24
 
25
- The CA-PT Parallel Corpus is a Catalan-Portuguese dataset of **9.892.953** parallel sentences. The dataset was created to support Catalan in NLP tasks, specifically
26
  Machine Translation.
27
 
28
 
@@ -39,11 +39,9 @@ The sentences included in the dataset are in Catalan (CA) and Portuguese (PT).
39
 
40
  ### Data Instances
41
 
42
- Two separate txt files are provided with the sentences sorted in the same order:
43
-
44
- - ca-pt_2023_09_01_full.ca: contains 9.892.953 Catalan sentences.
45
-
46
- - ca-pt_2023_09_01_full.pt: contains 9.892.953 Portuguese sentences.
47
 
48
 
49
  ### Data Fields
@@ -63,29 +61,19 @@ This dataset is aimed at promoting the development of Machine Translation betwee
63
  ### Source Data
64
 
65
  #### Initial Data Collection and Normalization
 
 
 
66
 
67
- The dataset is a combination of the following original datasets:
68
-
69
- | Dataset | Sentences |
70
- |:-------|-------:|
71
- | CCMatrix v1 | 3.765.459 |
72
- | WikiMatrix | 317.649 |
73
- | GNOME | 1.752 |
74
- | KDE4 | 117.828 |
75
- | OpenSubtitles | 235.604 |
76
- | GlobalVoices | 3.430 |
77
- | Tatoeba | 723 |
78
- | Europarl | 1.631.989 |
79
 
80
- All corpora except Europarl were collected from [Opus](https://opus.nlpl.eu/).
81
- The Europarl corpus is a synthetic parallel corpus created from the original Spanish-Catalan corpus by [SoftCatalà](https://github.com/Softcatala/Europarl-catalan).
82
-
83
- The remaining **3.733.322** sentences are synthetic parallel data created from a random sampling of the Spanish-Portuguese corpora
84
  available on [Opus](https://opus.nlpl.eu/) and translated into Catalan using the [PlanTL es-ca](https://huggingface.co/PlanTL-GOB-ES/mt-plantl-es-ca) model.
85
 
86
  All datasets are deduplicated and filtered to remove any sentence pairs with a cosine similarity of less than 0.75.
87
  This is done using sentence embeddings calculated using [LaBSE](https://huggingface.co/sentence-transformers/LaBSE).
88
- The filtered datasets are then concatenated to form a final corpus of **9.892.953** parallel sentences.
89
 
90
  #### Who are the source language producers?
91
 
@@ -102,7 +90,7 @@ The dataset does not contain any annotations.
102
  #### Who are the annotators?
103
 
104
  [N/A]
105
-
106
  ### Personal and Sensitive Information
107
 
108
  Given that this dataset is partly derived from pre-existing datasets that may contain crawled data, and that no specific anonymisation process has been applied,
 
22
 
23
  ### Dataset Summary
24
 
25
+ The CA-PT Parallel Corpus is a Catalan-Portuguese dataset of created to support Catalan in NLP tasks, specifically
26
  Machine Translation.
27
 
28
 
 
39
 
40
  ### Data Instances
41
 
42
+ A single tsv file is provided with the sentences sorted in the same order and
43
+ a header containing the two-letter ISO language code for the language in each column:
44
+ ca-pt_2023_09_01_full.tsv.
 
 
45
 
46
 
47
  ### Data Fields
 
61
  ### Source Data
62
 
63
  #### Initial Data Collection and Normalization
64
+ @alogin4 dataset_to_hf]$ /bin/python /gpfs/scratch/bsc88/bsc088063/dataset_to_hf/txt2parq
65
+ The first portion of the corpus is a combination of the following original datasets collected from [Opus](https://opus.nlpl.eu/):
66
+ CCMatrix, WikiMatrix, GNOME, KDE4, OpenSubtitles, GlobalVoices, Tatoeba.
67
 
68
+ Additionally, the corpus contains synthetic parallel data generated from the original Spanish-Catalan Europarl
69
+ made public by [SoftCatalà](https://github.com/Softcatala/Europarl-catalan).
 
 
 
 
 
 
 
 
 
 
70
 
71
+ A last portion of the dataset is composed by synthetic parallel data generated from a random sampling of the Spanish-Portuguese corpora
 
 
 
72
  available on [Opus](https://opus.nlpl.eu/) and translated into Catalan using the [PlanTL es-ca](https://huggingface.co/PlanTL-GOB-ES/mt-plantl-es-ca) model.
73
 
74
  All datasets are deduplicated and filtered to remove any sentence pairs with a cosine similarity of less than 0.75.
75
  This is done using sentence embeddings calculated using [LaBSE](https://huggingface.co/sentence-transformers/LaBSE).
76
+ The filtered datasets are then concatenated to form the final corpus.
77
 
78
  #### Who are the source language producers?
79
 
 
90
  #### Who are the annotators?
91
 
92
  [N/A]
93
+ @alogin4 dataset_to_hf]$ /bin/python /gpfs/scratch/bsc88/bsc088063/dataset_to_hf/txt2parq
94
  ### Personal and Sensitive Information
95
 
96
  Given that this dataset is partly derived from pre-existing datasets that may contain crawled data, and that no specific anonymisation process has been applied,