Hugging Face's logo

Datasets:
techiaith
/
commonvoice_23_0_en__GB_IE

Modalities:
Audio
Text
Formats:
parquet
Size:
100K - 1M
Libraries:
Datasets
Dask
Dataset card Data Studio Files
xet
Community
1

You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this dataset content.

Common Voice 23.0: English Language Subset for Great Britain and Ireland

Split Duration Word Count No Of Clips
train 224:58:09 1,181,302 164,404
dev 26:55:57 140,431 19,798
test 13:17:05 70,365 9732

Split percentages: Train=85%, Dev=10%, Test=5%
Target accents: Welsh English, England English, Scottish English, Northern Irish, Irish English

GB/IE accented recordings from Common Voice 23 (en) validated.tsv...
Total validated recordings: 1,822,716
GB/IE accented recordings: 193,934

Breakdown by accent:

Welsh English: 2,103
England English: 153,337
Scottish English: 18,718
Northern Irish: 10,482
Irish English: 9,294

CREATING SPLITS WITH SENTENCE EXCLUSIVITY

Total unique sentences: 143,011
Total recordings: 193,934
Total unique speakers: 3,213

Sentence distribution:

  • Train: 121,559 sentences (85.0%)
  • Dev: 14,301 sentences (10.0%)
  • Test: 7,151 sentences (5.0%)

Final recording distribution:

  • Train: 164,404 recordings (84.8%)
  • Dev: 19,798 recordings (10.2%)
  • Test: 9,732 recordings (5.0%)

Speaker distribution: (speakers MAY appear in multiple splits)

  • Train: 3,162 speakers
  • Dev: 2,106 speakers
  • Test: 1,635 speakers

Speaker overlap:

  • Train-Dev: 2,072 speakers
  • Train-Test: 1,615 speakers
  • Dev-Test: 1,370 speakers
Downloads last month
3

Collection including techiaith/commonvoice_23_0_en__GB_IE