ingerid commited on
Commit
e47cee1
1 Parent(s): 8e3b1c0

docs: update dataset card.

Browse files

Change split name 'dev' to 'validation'

Credit the huggingface team for help

Files changed (1) hide show
  1. README.md +3 -75
README.md CHANGED
@@ -12,78 +12,6 @@ tags:
12
  - live-events
13
  - conversational
14
  - speech
15
- dataset_info:
16
- features:
17
- - name: source_file_id
18
- dtype: string
19
- - name: segment_id
20
- dtype: string
21
- - name: segment_order
22
- dtype: int64
23
- - name: duration
24
- dtype: float64
25
- - name: overlap_previous
26
- dtype: bool
27
- - name: overlap_next
28
- dtype: bool
29
- - name: speaker_id
30
- dtype: string
31
- - name: gender
32
- dtype:
33
- class_label:
34
- names:
35
- '0': f
36
- '1': m
37
- - name: dialect
38
- dtype:
39
- class_label:
40
- names:
41
- '0': e
42
- '1': n
43
- '2': sw
44
- '3': t
45
- '4': w
46
- - name: orthography
47
- dtype:
48
- class_label:
49
- names:
50
- '0': bm
51
- '1': nn
52
- - name: source_type
53
- dtype:
54
- class_label:
55
- names:
56
- '0': live-event
57
- '1': podcast
58
- - name: file_name
59
- dtype: string
60
- - name: transcription
61
- dtype: string
62
- - name: audio
63
- dtype:
64
- audio:
65
- sampling_rate: 16000
66
- splits:
67
- - name: dev
68
- num_bytes: 261655958.008
69
- num_examples: 1194
70
- - name: train
71
- num_bytes: 2010500512.784
72
- num_examples: 9548
73
- - name: test
74
- num_bytes: 268382590.555
75
- num_examples: 1195
76
- download_size: 2731120894
77
- dataset_size: 2540539061.3469996
78
- configs:
79
- - config_name: default
80
- data_files:
81
- - split: dev
82
- path: data/dev-*
83
- - split: train
84
- path: data/train-*
85
- - split: test
86
- path: data/test-*
87
  ---
88
 
89
  # Dataset Card for Sprakbanken/nb_samtale
@@ -174,7 +102,7 @@ data field | description | Value type / example
174
 
175
  ### Data Splits
176
 
177
- The data is split into a training, development, and test set, stratified on three parameters: source type, gender and dialect.
178
  Gender and dialect naturally refers to the gender and dialect of the speakers.
179
  The data has not been split on speaker ID to avoid speaker overlap in the various sets because this proved impossible
180
  while still maintaining a decent distribution of the other parameters, especially dialect variation.
@@ -242,8 +170,8 @@ speaker time, while the other areas fall between 8% and 20% speaker time.
242
  ### Dataset Curators
243
 
244
  The content of the dataset was created by the Norwegian Language Bank (Språkbanken) at the National Library of Norway.
245
- [Marie Iversdatter Røsok](mailto:[email protected]), [Ingerid Løyning Dale](mailto:[email protected]) and [Per Erik Solberg](mailto:[email protected]) all contributed in creating this dataset.
246
- <!-- Thanks to the HuggingFace team for assistance. -->
247
 
248
  ### Licensing Information
249
 
 
12
  - live-events
13
  - conversational
14
  - speech
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
15
  ---
16
 
17
  # Dataset Card for Sprakbanken/nb_samtale
 
102
 
103
  ### Data Splits
104
 
105
+ The data is split into a `train`, `validation`, and `test` set, stratified on three parameters: source type, gender and dialect.
106
  Gender and dialect naturally refers to the gender and dialect of the speakers.
107
  The data has not been split on speaker ID to avoid speaker overlap in the various sets because this proved impossible
108
  while still maintaining a decent distribution of the other parameters, especially dialect variation.
 
170
  ### Dataset Curators
171
 
172
  The content of the dataset was created by the Norwegian Language Bank (Språkbanken) at the National Library of Norway.
173
+ [Marie Iversdatter Røsok](mailto:[email protected]), [Ingerid Løyning Dale](mailto:[email protected]) and [Per Erik Solberg](mailto:[email protected]) contributed in creating this dataset.
174
+ Thanks to the HuggingFace team for assistance.
175
 
176
  ### Licensing Information
177