Baybars commited on
Commit
280139a
1 Parent(s): cff1573

Update README.md

Browse files

model card additions

Files changed (1) hide show
  1. README.md +57 -12
README.md CHANGED
@@ -57,8 +57,8 @@ task_ids:
57
 
58
  - **Homepage:** https://zenodo.org/record/5541827
59
  - **Repository:** https://github.com/CollectivaT-dev/ParlamentParla
60
- - **Paper:**
61
- - **Point of Contact:** [Col·lectivaT](mailto:info@collectivat.cat)
62
 
63
  ### Dataset Summary
64
 
@@ -118,53 +118,86 @@ The dataset is split in: "train", "validation" and "test".
118
 
119
  ## Dataset Creation
120
 
 
 
 
 
121
  ### Curation Rationale
122
 
123
- [More Information Needed]
124
 
125
  ### Source Data
126
 
127
  #### Initial Data Collection and Normalization
128
 
129
- [More Information Needed]
 
 
 
130
 
131
  #### Who are the source language producers?
132
 
133
- [More Information Needed]
 
134
 
135
  ### Annotations
136
 
 
 
137
  #### Annotation process
138
 
139
- [More Information Needed]
140
 
141
  #### Who are the annotators?
142
 
143
- [More Information Needed]
144
 
145
  ### Personal and Sensitive Information
146
 
147
- [More Information Needed]
 
148
 
149
  ## Considerations for Using the Data
150
 
151
  ### Social Impact of Dataset
152
 
153
- [More Information Needed]
 
154
 
155
  ### Discussion of Biases
156
 
157
- [More Information Needed]
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
158
 
159
  ### Other Known Limitations
160
 
161
- [More Information Needed]
162
 
163
  ## Additional Information
164
 
165
  ### Dataset Curators
166
 
167
- [More Information Needed]
 
168
 
169
  ### Licensing Information
170
 
@@ -185,6 +218,18 @@ The dataset is split in: "train", "validation" and "test".
185
  url = {https://doi.org/10.5281/zenodo.5541827}
186
  }
187
  ```
 
 
 
 
 
 
 
 
 
 
 
 
188
  ### Funding
189
 
190
  This work was funded by the [Catalan Ministry of the Vice-presidency, Digital Policies and Territory](https://politiquesdigitals.gencat.cat/en/inici/index.html) within the framework of the [Aina project](https://politiquesdigitals.gencat.cat/ca/tic/aina-el-projecte-per-garantir-el-catala-en-lera-digital/).
 
57
 
58
  - **Homepage:** https://zenodo.org/record/5541827
59
  - **Repository:** https://github.com/CollectivaT-dev/ParlamentParla
60
+ - **Paper:** ParlamentParla: [A Speech Corpus of Catalan Parliamentary Sessions.](http://www.lrec-conf.org/proceedings/lrec2022/workshops/ParlaCLARINIII/2022.parlaclariniii-1.0.pdf#page=135)
61
+ - **Point of Contact:** [Col·lectivaT](mailto:baybars.kulebi@bsc.es)
62
 
63
  ### Dataset Summary
64
 
 
118
 
119
  ## Dataset Creation
120
 
121
+ The dataset is created by aligning the parliamentary session transcripts
122
+ and the audiovisual content. For more detailed information please consult
123
+ this [paper](http://www.lrec-conf.org/proceedings/lrec2022/workshops/ParlaCLARINIII/2022.parlaclariniii-1.0.pdf#page=135).
124
+
125
  ### Curation Rationale
126
 
127
+ We created this corpus to contribute to the development of language models in Catalan, a low-resource language.
128
 
129
  ### Source Data
130
 
131
  #### Initial Data Collection and Normalization
132
 
133
+ The audio segments were extracted from recordings the Catalan Parliament
134
+ (Parlament de Catalunya) plenary sessions, which took place between 2007/07/11 -
135
+ 2018/07/17. The cleaning procedures are in the archived repository [Long Audio
136
+ Aligner](https://github.com/gullabi/long-audio-aligner)
137
 
138
  #### Who are the source language producers?
139
 
140
+ The parliamentary members of the legislatures between 2007/07/11 -
141
+ 2018/07/17
142
 
143
  ### Annotations
144
 
145
+ The dataset is unannotated.
146
+
147
  #### Annotation process
148
 
149
+ [N/A]
150
 
151
  #### Who are the annotators?
152
 
153
+ [N/A]
154
 
155
  ### Personal and Sensitive Information
156
 
157
+ The initial content is publicly available furthermore, the identities of
158
+ the parliamentary members are anonymized.
159
 
160
  ## Considerations for Using the Data
161
 
162
  ### Social Impact of Dataset
163
 
164
+ We hope this corpus contributes to the development of language models in
165
+ Catalan, a low-resource language.
166
 
167
  ### Discussion of Biases
168
 
169
+ This dataset has a gender bias, however since the speakers are tagged according to
170
+ their genders, creating a balanced subcorpus is possible.
171
+
172
+ | Subcorpus | Gender | Duration (h) |
173
+ |-------------|----------|------------|
174
+ | other_test | F | 2.516 |
175
+ | other_dev | F | 2.701 |
176
+ | other_train | F | 109.68 |
177
+ | other_test | M | 2.631 |
178
+ | other_dev | M | 2.513 |
179
+ | other_train | M | 280.196 |
180
+ |*other total*| | 400.239 |
181
+ | clean_test | F | 2.707 |
182
+ | clean_dev | F | 2.576 |
183
+ | clean_train | F | 77.905 |
184
+ | clean_test | M | 2.516 |
185
+ | clean_dev | M | 2.614 |
186
+ | clean_train | M | 123.162 |
187
+ |*clean total*| | 211.48 |
188
+ |*Total* | | 611.719 |
189
+
190
 
191
  ### Other Known Limitations
192
 
193
+ The text corpus belongs to the domain of Catalan politics
194
 
195
  ## Additional Information
196
 
197
  ### Dataset Curators
198
 
199
+ Baybars Külebi ([@gullabi](https://github.com/gullabi),
200
+ [@baybars](https://huggingface.co/baybars))
201
 
202
  ### Licensing Information
203
 
 
218
  url = {https://doi.org/10.5281/zenodo.5541827}
219
  }
220
  ```
221
+
222
+ For the paper:
223
+ @inproceedings{kulebi2022parlamentparla,
224
+ title={ParlamentParla: A Speech Corpus of Catalan Parliamentary Sessions},
225
+ author={K{\"u}lebi, Baybars and Armentano-Oller, Carme and Rodr{\'\i}guez-Penagos, Carlos and Villegas, Marta},
226
+ booktitle={Workshop on Creating, Enriching and Using Parliamentary Corpora},
227
+ volume={125},
228
+ number={130},
229
+ pages={125},
230
+ year={2022}
231
+ }
232
+
233
  ### Funding
234
 
235
  This work was funded by the [Catalan Ministry of the Vice-presidency, Digital Policies and Territory](https://politiquesdigitals.gencat.cat/en/inici/index.html) within the framework of the [Aina project](https://politiquesdigitals.gencat.cat/ca/tic/aina-el-projecte-per-garantir-el-catala-en-lera-digital/).