spapi commited on
Commit
f1f0f39
1 Parent(s): 0f884d2

Update metadata and dataset visualization (#1)

Browse files

- Update metadata and dataset visualization (36c14697333f9b34d8e1f63bd472205efc627bf0)
- Update splits name with original dataset specifications (94862e5aea79d3094cadd725934ec795545f6616)
- Change path of train_voxpopuli splits (e05028c2634135a98761e86ddc1e2adfed76083c)
- Update arXiv link (48c5e617c185ee4e0c45e29099251dcf5916ecfd)

Files changed (1) hide show
  1. README.md +231 -121
README.md CHANGED
@@ -1,121 +1,231 @@
1
- ---
2
- task_categories:
3
- - automatic-speech-recognition
4
- language:
5
- - en
6
- - bg
7
- - hr
8
- - cs
9
- - da
10
- - nl
11
- - et
12
- - fi
13
- - fr
14
- - de
15
- - el
16
- - hu
17
- - ga
18
- - it
19
- - lv
20
- - lt
21
- - mt
22
- - pl
23
- - pt
24
- - ro
25
- - sk
26
- - sl
27
- - es
28
- - sv
29
- pretty_name: MOSEL
30
- license: cc-by-4.0
31
- ---
32
-
33
- <img src="./mosel-logo-transparent.png" align="center" width="100%">
34
-
35
- ### Dataset Description, Collection, and Source
36
-
37
- The MOSEL corpus is a multilingual dataset collection including up to 950K hours of open-source speech recordings covering the 24 official languages of the European Union. We collect data by surveying labeled and unlabeled speech corpora under open-source compliant licenses.
38
- In particular, MOSEL includes the automatic transcripts of 441k hours of unlabeled speech from VoxPopuli and LibriLight. The data is transcribed using [Whisper large v3](https://huggingface.co/openai/whisper-large-v3).
39
- Whisper is released under the OS Apache 2.0 License which allows releasing the generated content under any license. Since LibriLight, differently from VoxPopuli, contains segments longer than Whisper's maximum duration limit of 30sec, we split them into chunks of up to 30sec.
40
-
41
- - **Curated by:** Marco Gaido, Sara Papi, Luisa Bentivogli, Alessio Brutti, Mauro Cettolo, Roberto Gretter, Marco Matassoni, Mohamed Nabih, and Matteo Negri
42
- - **Funded by:** FAIR, Meetween, and CINECA
43
- - **Shared by:** Fondazione Bruno Kessler
44
-
45
- ### License
46
- - CC-BY-4.0
47
-
48
- ### Dataset Sources
49
-
50
- - **Collection Repository:** [MOSEL](https://github.com/hlt-mt/mosel)
51
- - **Paper:** [MOSEL: 950,000 Hours of Speech Data for Open-Source Speech Foundation Model Training on EU Languages](https://arxiv.org/)
52
-
53
- ## Dataset Structure
54
-
55
- ### Data Config
56
- The dataset is split into folders corresponding to the languages using the [2-letters ISO codes](https://en.wikipedia.org/wiki/List_of_ISO_639_language_codes), one for each language. Within each folder, a split for each psuedo-labeled dataset is provided.
57
-
58
- ### Data Field
59
- `id`: alphanumeric identifier for the segment
60
-
61
- `language`: extended language (e.g., "english")
62
-
63
- `text`: the content of the psuedo label
64
-
65
- `hall_repeated_ngrams`: True/False - indicates the repetition of an *n*-gram in `text` for a minimum number of times; for *n* in 1 to 2, the threshold is 4, for *n* in 3 to 5, it is 3
66
-
67
- `hall_long_word`: True/False - indicates the presence of a word of at least 40 characters in `text`
68
-
69
- `hall_frequent_single_word`: True/False - indicates that `text` consists of only one word which is the most frequent inside the whole text
70
-
71
- ## Dataset Statistics (in hours)
72
-
73
- | Language (LangID) | Labeled | Unlabeled | Total |
74
- |--------|--------|--------|-------|
75
- | Bulgarian (bg) | 111 | 17609 | 17720 |
76
- | Croatian (hr) | 55 | 8106 | 8161 |
77
- | Czech (cs) | 591 | 18705 | 19296 |
78
- | Danish (da) | 20 | 13600 | 13620 |
79
- | Dutch (nl) | 3395 | 19014 | 22409 |
80
- | English (en) | 437239 | 84704 | 521943|
81
- | Estonian (et) | 60 | 10604 | 10664 |
82
- | Finnish (fi) | 64 | 14200 | 14264 |
83
- | French (fr) | 26984 | 22896 | 49880 |
84
- | German (de) | 9236 | 23228 | 32464 |
85
- | Greek (el) | 35 | 17703 | 17738 |
86
- | Hungarian (hu) | 189 | 17701 | 17890 |
87
- | Irish (ga) | 17 | 0 | 17 |
88
- | Italian (it) | 3756 | 21933 | 25689 |
89
- | Latvian (lv) | 173 | 13100 | 13273 |
90
- | Lithuanian (lt) | 36 | 14400 | 14436 |
91
- | Maltese (mt) | 19 | 9100 | 9119 |
92
- | Polish (pl) | 510 | 21207 | 21717 |
93
- | Portuguese (pt) | 5492 | 17526 | 23018 |
94
- | Romanian (ro) | 121 | 17906 | 18021 |
95
- | Slovak (sk) | 61 | 12100 | 12161 |
96
- | Slovenian (sl) | 32 | 11300 | 11332 |
97
- | Spanish (es) | 17471 | 21526 | 38997 |
98
- | Swedish (sv) | 58 | 16300 | 16358 |
99
- | Total | 505725 | 444467 | 950192|
100
-
101
-
102
- ## Dataset Creation
103
- To reproduce the dataset creation, please refer to the [MOSEL README in the fbk-llm](https://github.com/hlt-mt/fbk-llm) repository.
104
-
105
-
106
- ## Citation
107
- Release 1.0:
108
- ```
109
- @inproceedings{mosel,
110
- title = {{MOSEL: 950,000 Hours of Speech Data for Open-Source Speech Foundation Model Training on EU Languages}},
111
- author = {Marco Gaido and Sara Papi and Luisa Bentivogli and Alessio Brutti and Mauro Cettolo and Roberto Gretter and Marco Matassoni and Mohamed Nabihand Matteo Negri},
112
- booktitle = "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing",
113
- month = nov,
114
- year = "2024",
115
- address = "Miami, United States",
116
- publisher = "Association for Computational Linguistics",
117
- }
118
- ```
119
-
120
- ## Dataset Card Contact
121
- [@spapi](https://huggingface.co/spapi)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ annotations_creators:
3
+ - machine-generated
4
+ task_categories:
5
+ - automatic-speech-recognition
6
+ - text-to-speech
7
+ language:
8
+ - en
9
+ - bg
10
+ - hr
11
+ - cs
12
+ - da
13
+ - nl
14
+ - et
15
+ - fi
16
+ - fr
17
+ - de
18
+ - el
19
+ - hu
20
+ - ga
21
+ - it
22
+ - lv
23
+ - lt
24
+ - mt
25
+ - pl
26
+ - pt
27
+ - ro
28
+ - sk
29
+ - sl
30
+ - es
31
+ - sv
32
+ language_creators:
33
+ - found
34
+ modality:
35
+ - text
36
+ - audio
37
+ multilinguality:
38
+ - multilingual
39
+ pretty_name: MOSEL
40
+ license: cc-by-4.0
41
+ tags:
42
+ - speech
43
+ - speech-to-text
44
+ - open-source
45
+ - whisper
46
+ configs:
47
+ - config_name: bg
48
+ data_files:
49
+ - split: train_voxpopuli
50
+ path: bg/voxpopuli*
51
+ - config_name: cs
52
+ data_files:
53
+ - split: train_voxpopuli
54
+ path: cs/voxpopuli*
55
+ - config_name: da
56
+ data_files:
57
+ - split: train_voxpopuli
58
+ path: da/voxpopuli*
59
+ - config_name: de
60
+ data_files:
61
+ - split: train_voxpopuli
62
+ path: de/voxpopuli*
63
+ - config_name: el
64
+ data_files:
65
+ - split: train_voxpopuli
66
+ path: el/voxpopuli*
67
+ - config_name: en
68
+ data_files:
69
+ - split: train_voxpopuli
70
+ path: en/voxpopuli*
71
+ - split: train_librilight
72
+ path: en/librilight*
73
+ - config_name: es
74
+ data_files:
75
+ - split: train_voxpopuli
76
+ path: es/voxpopuli*
77
+ - config_name: et
78
+ data_files:
79
+ - split: train_voxpopuli
80
+ path: et/voxpopuli*
81
+ - config_name: fi
82
+ data_files:
83
+ - split: train_voxpopuli
84
+ path: fi/voxpopuli*
85
+ - config_name: fr
86
+ data_files:
87
+ - split: train_voxpopuli
88
+ path: fr/voxpopuli*
89
+ - config_name: hr
90
+ data_files:
91
+ - split: train_voxpopuli
92
+ path: hr/voxpopuli*
93
+ - config_name: hu
94
+ data_files:
95
+ - split: train_voxpopuli
96
+ path: hu/voxpopuli*
97
+ - config_name: it
98
+ data_files:
99
+ - split: train_voxpopuli
100
+ path: it/voxpopuli*
101
+ - config_name: lt
102
+ data_files:
103
+ - split: train_voxpopuli
104
+ path: lt/voxpopuli*
105
+ - config_name: lv
106
+ data_files:
107
+ - split: train_voxpopuli
108
+ path: lv/voxpopuli*
109
+ - config_name: mt
110
+ data_files:
111
+ - split: train_voxpopuli
112
+ path: mt/voxpopuli*
113
+ - config_name: nl
114
+ data_files:
115
+ - split: train_voxpopuli
116
+ path: nl/voxpopuli*
117
+ - config_name: pl
118
+ data_files:
119
+ - split: train_voxpopuli
120
+ path: pl/voxpopuli*
121
+ - config_name: pt
122
+ data_files:
123
+ - split: train_voxpopuli
124
+ path: pt/voxpopuli*
125
+ - config_name: ro
126
+ data_files:
127
+ - split: train_voxpopuli
128
+ path: ro/voxpopuli*
129
+ - config_name: sk
130
+ data_files:
131
+ - split: train_voxpopuli
132
+ path: sk/voxpopuli*
133
+ - config_name: sl
134
+ data_files:
135
+ - split: train_voxpopuli
136
+ path: sl/voxpopuli*
137
+ - config_name: sv
138
+ data_files:
139
+ - split: train_voxpopuli
140
+ path: sv/voxpopuli*
141
+ ---
142
+
143
+ <img src="./mosel-logo-transparent.png" align="center" width="100%">
144
+
145
+ ### Dataset Description, Collection, and Source
146
+
147
+ The MOSEL corpus is a multilingual dataset collection including up to 950K hours of open-source speech recordings covering the 24 official languages of the European Union. We collect data by surveying labeled and unlabeled speech corpora under open-source compliant licenses.
148
+ In particular, MOSEL includes the automatic transcripts of 441k hours of unlabeled speech from VoxPopuli and LibriLight. The data is transcribed using [Whisper large v3](https://huggingface.co/openai/whisper-large-v3).
149
+ Whisper is released under the OS Apache 2.0 License which allows releasing the generated content under any license. Since LibriLight, differently from VoxPopuli, contains segments longer than Whisper's maximum duration limit of 30sec, we split them into chunks of up to 30sec.
150
+
151
+ - **Curated by:** Marco Gaido, Sara Papi, Luisa Bentivogli, Alessio Brutti, Mauro Cettolo, Roberto Gretter, Marco Matassoni, Mohamed Nabih, and Matteo Negri
152
+ - **Funded by:** FAIR, Meetween, and CINECA
153
+ - **Shared by:** Fondazione Bruno Kessler
154
+
155
+ ### License
156
+ - CC-BY-4.0
157
+
158
+ ### Dataset Sources
159
+
160
+ - **Collection Repository:** [MOSEL](https://github.com/hlt-mt/mosel)
161
+ - **Paper:** [MOSEL: 950,000 Hours of Speech Data for Open-Source Speech Foundation Model Training on EU Languages](http://arxiv.org/abs/2410.01036)
162
+
163
+ ## Dataset Structure
164
+
165
+ ### Data Config
166
+ The dataset is split into folders corresponding to the languages using the [2-letters ISO codes](https://en.wikipedia.org/wiki/List_of_ISO_639_language_codes), one for each language. Within each folder, a split for each psuedo-labeled dataset is provided.
167
+
168
+ ### Data Field
169
+ `id`: alphanumeric identifier for the segment
170
+
171
+ `language`: extended language (e.g., "english")
172
+
173
+ `text`: the content of the psuedo label
174
+
175
+ `hall_repeated_ngrams`: True/False - indicates the repetition of an *n*-gram in `text` for a minimum number of times; for *n* in 1 to 2, the threshold is 4, for *n* in 3 to 5, it is 3
176
+
177
+ `hall_long_word`: True/False - indicates the presence of a word of at least 40 characters in `text`
178
+
179
+ `hall_frequent_single_word`: True/False - indicates that `text` consists of only one word which is the most frequent inside the whole text
180
+
181
+ ## Dataset Statistics (in hours)
182
+
183
+ | Language (LangID) | Labeled | Unlabeled | Total |
184
+ |--------|--------|--------|-------|
185
+ | Bulgarian (bg) | 111 | 17609 | 17720 |
186
+ | Croatian (hr) | 55 | 8106 | 8161 |
187
+ | Czech (cs) | 591 | 18705 | 19296 |
188
+ | Danish (da) | 20 | 13600 | 13620 |
189
+ | Dutch (nl) | 3395 | 19014 | 22409 |
190
+ | English (en) | 437239 | 84704 | 521943|
191
+ | Estonian (et) | 60 | 10604 | 10664 |
192
+ | Finnish (fi) | 64 | 14200 | 14264 |
193
+ | French (fr) | 26984 | 22896 | 49880 |
194
+ | German (de) | 9236 | 23228 | 32464 |
195
+ | Greek (el) | 35 | 17703 | 17738 |
196
+ | Hungarian (hu) | 189 | 17701 | 17890 |
197
+ | Irish (ga) | 17 | 0 | 17 |
198
+ | Italian (it) | 3756 | 21933 | 25689 |
199
+ | Latvian (lv) | 173 | 13100 | 13273 |
200
+ | Lithuanian (lt) | 36 | 14400 | 14436 |
201
+ | Maltese (mt) | 19 | 9100 | 9119 |
202
+ | Polish (pl) | 510 | 21207 | 21717 |
203
+ | Portuguese (pt) | 5492 | 17526 | 23018 |
204
+ | Romanian (ro) | 121 | 17906 | 18021 |
205
+ | Slovak (sk) | 61 | 12100 | 12161 |
206
+ | Slovenian (sl) | 32 | 11300 | 11332 |
207
+ | Spanish (es) | 17471 | 21526 | 38997 |
208
+ | Swedish (sv) | 58 | 16300 | 16358 |
209
+ | Total | 505725 | 444467 | 950192|
210
+
211
+
212
+ ## Dataset Creation
213
+ To reproduce the dataset creation, please refer to the [MOSEL README in the fbk-llm](https://github.com/hlt-mt/fbk-llm) repository.
214
+
215
+
216
+ ## Citation
217
+ Release 1.0:
218
+ ```
219
+ @inproceedings{mosel,
220
+ title = {{MOSEL: 950,000 Hours of Speech Data for Open-Source Speech Foundation Model Training on EU Languages}},
221
+ author = {Marco Gaido and Sara Papi and Luisa Bentivogli and Alessio Brutti and Mauro Cettolo and Roberto Gretter and Marco Matassoni and Mohamed Nabihand Matteo Negri},
222
+ booktitle = "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing",
223
+ month = nov,
224
+ year = "2024",
225
+ address = "Miami, United States",
226
+ publisher = "Association for Computational Linguistics",
227
+ }
228
+ ```
229
+
230
+ ## Dataset Card Contact
231
+ [@spapi](https://huggingface.co/spapi)