Datasets:

Modalities:
Text
Formats:
parquet
DOI:
Libraries:
Datasets
pandas
License:
fdelucaf commited on
Commit
89d2129
·
1 Parent(s): 48cd82d

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +24 -19
README.md CHANGED
@@ -59,9 +59,9 @@ The texts in the dataset are in Catalan and Portuguese.
59
 
60
  Two separated txt files are provided with the sentences sorted in the same order:
61
 
62
- - ca-pt_2023_09_01_full.ca: contains the Catalan sentences
63
 
64
- - ca-pt_2023_09_01_full.pt: contains the Portuguese sentences
65
 
66
  ### Data Splits
67
 
@@ -73,25 +73,29 @@ The dataset contains a single split: `train`.
73
 
74
  The dataset is a combination of the following datasets:
75
 
76
- | Dataset | Sentences | Sentences after Cleaning|
77
- |-------------------|----------------|-------------------|
78
- | CCMatrix v1 | 12.674.684 | 3.765.459|
79
- | WikiMatrix | 358.873 | 317.649 |
80
- | GNOME | 5.211 | 1.752|
81
- | KDE4 | 166.208 | 117.828 |
82
- | QED | 53.635 | 43.736 |
83
- | TED2020 v1 | 48.942 | 41.461 |
84
- | OpenSubtitles | 384.142 | 235.604 |
85
- | GlobalVoices| 4.035 | 3.430|
86
- | Tatoeba | 754 | 723 |
87
- | Europarl | 1.692.106 | 3.765.459 |
88
- | **Total** | **15.391.745** | **6.159.631** |
89
-
90
- All corpora except Europarl were collected from [Opus](https://opus.nlpl.eu/). The Europarl corpus is a synthetic parallel corpus created from the original Spanish-Catalan corpus by [SoftCatalà](https://github.com/Softcatala/Europarl-catalan).
 
 
91
 
92
  ### Data preparation
93
 
94
- All datasets are deduplicated and filtered to remove any sentence pairs with a cosine similarity of less than 0.75. This is done using sentence embeddings calculated using [LaBSE](https://huggingface.co/sentence-transformers/LaBSE). The filtered datasets are then concatenated to form a final corpus of 6.159.631 and before training the punctuation is normalized using a modified version of the join-single-file.py script from [SoftCatalà](https://github.com/Softcatala/nmt-models/blob/master/data-processing-tools/join-single-file.py)
 
 
95
 
96
  ### Personal and Sensitive Information
97
 
@@ -105,7 +109,8 @@ The purpose of this dataset is to help develop Machines Translation tasks for lo
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. Nonetheless, we have not applied any steps to reduce their impact.
 
109
 
110
  ### Other Known Limitations
111
 
 
59
 
60
  Two separated txt files are provided with the sentences sorted in the same order:
61
 
62
+ - ca-pt_2023_09_01_full.ca: contains 9 892 953 Catalan sentences
63
 
64
+ - ca-pt_2023_09_01_full.pt: contains 9 892 953 Portuguese sentences
65
 
66
  ### Data Splits
67
 
 
73
 
74
  The dataset is a combination of the following datasets:
75
 
76
+ | Dataset | Sentences |
77
+ |-------------------|-----------|
78
+ | CCMatrix v1 | 3.765.459 |
79
+ | WikiMatrix | 317.649 |
80
+ | GNOME | 1.752 |
81
+ | KDE4 | 117.828 |
82
+ | QED | 43.736 |
83
+ | TED2020 v1 | 41.461 |
84
+ | OpenSubtitles | 235.604 |
85
+ | GlobalVoices | 3.430 |
86
+ | Tatoeba | 723 |
87
+ | Europarl | 3.765.459 |
88
+ | **Total** | **6.159.631** |
89
+
90
+ All corpora except Europarl were collected from [Opus](https://opus.nlpl.eu/).
91
+ The Europarl corpus is a synthetic parallel corpus created from the original Spanish-Catalan corpus by [SoftCatalà](https://github.com/Softcatala/Europarl-catalan).
92
+
93
 
94
  ### Data preparation
95
 
96
+ All datasets are deduplicated and filtered to remove any sentence pairs with a cosine similarity of less than 0.75.
97
+ This is done using sentence embeddings calculated using [LaBSE](https://huggingface.co/sentence-transformers/LaBSE).
98
+ The filtered datasets are then concatenated to form a final corpus of 6.159.631 and before training the punctuation is normalized using a modified version of the join-single-file.py script from [SoftCatalà](https://github.com/Softcatala/nmt-models/blob/master/data-processing-tools/join-single-file.py)
99
 
100
  ### Personal and Sensitive Information
101
 
 
109
 
110
  ### Discussion of Biases
111
 
112
+ 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.
113
+ Nonetheless, we have not applied any steps to reduce their impact.
114
 
115
  ### Other Known Limitations
116