Hugging Face's logo

Datasets:
SoundWaveET
/
leyu-amharic-speech-corpus-2026

Tasks:
Automatic Speech Recognition
Modalities:
Audio
Languages:
Amharic
Oriya
Tigrinya
Size:
1K<n<10K
Tags:
speech
audio
ethiopian-languages
amharic
afaan-oromo
tigrinya
License:
Dataset card Files
xet
 Community
1
Dataset Viewer
Duplicate
The dataset viewer is not available for this split.
Cannot load the dataset split (in streaming mode) to extract the first rows.
Error code:   StreamingRowsError
Exception:    CastError
Message:      Couldn't cast
audio: string
file_name: string
audio_id: string
language: string
language_code: string
dialect: string
microtask_id: string
microtask_title: string
transcription: string
speaker_id: string
contributor_name: string
contributor_phone: int64
contributor_place: string
gender: string
duration: double
sampling_rate: int64
audio_format: string
-- schema metadata --
pandas: '{"index_columns": [{"kind": "range", "name": null, "start": 0, "' + 2351
to
{'audio': Audio(sampling_rate=None, decode=False, num_channels=None, stream_index=None), 'audio_id': Value('string'), 'transcription': Value('string'), 'language': Value('string'), 'language_code': Value('string'), 'dialect': Value('string'), 'microtask_id': Value('string'), 'microtask_title': Value('string'), 'speaker_id': Value('string'), 'contributor_name': Value('string'), 'contributor_phone': Value('string'), 'contributor_place': Value('string'), 'gender': Value('string'), 'duration': Value('float64'), 'sampling_rate': Value('int64'), 'audio_format': Value('string')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 147, in get_rows_or_raise
                  return get_rows(
                      dataset=dataset,
                  ...<4 lines>...
                      column_names=column_names,
                  )
                File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
                  return func(*args, **kwargs)
                File "/src/services/worker/src/worker/utils.py", line 127, in get_rows
                  rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
                File "/src/services/worker/src/worker/utils.py", line 483, in safe_iter
                  yield from ds.decode(False) if ds.features else ds
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2840, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2373, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2406, in _iter_arrow
                  pa_table = cast_table_to_features(pa_table, self.features)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2280, in cast_table_to_features
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              audio: string
              file_name: string
              audio_id: string
              language: string
              language_code: string
              dialect: string
              microtask_id: string
              microtask_title: string
              transcription: string
              speaker_id: string
              contributor_name: string
              contributor_phone: int64
              contributor_place: string
              gender: string
              duration: double
              sampling_rate: int64
              audio_format: string
              -- schema metadata --
              pandas: '{"index_columns": [{"kind": "range", "name": null, "start": 0, "' + 2351
              to
              {'audio': Audio(sampling_rate=None, decode=False, num_channels=None, stream_index=None), 'audio_id': Value('string'), 'transcription': Value('string'), 'language': Value('string'), 'language_code': Value('string'), 'dialect': Value('string'), 'microtask_id': Value('string'), 'microtask_title': Value('string'), 'speaker_id': Value('string'), 'contributor_name': Value('string'), 'contributor_phone': Value('string'), 'contributor_place': Value('string'), 'gender': Value('string'), 'duration': Value('float64'), 'sampling_rate': Value('int64'), 'audio_format': Value('string')}
              because column names don't match

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

🇪🇹 Leyu Ethiopian Multilingual Speech Corpus 2026

Official research-grade multilingual speech dataset submitted for the Leyu Platform Data Collection Competition 2026, organized by gheero (Leyu Platform Team).

🏢 Platform & Team Verification Details

🌟 Dataset Architecture & Standardized Naming Convention

Every audio sample follows the structured Leyu data engineering convention: audio/leyu_eth_<language_code>_<dialect_code>_<task_id>_<speaker_id>_<record_id>.<ext>

All audio durations strictly adhere to the competition rule (10 to 60 seconds per sample).

📊 Linguistic & Demographic Breakdown

Audio Identifier Language ISO Code Dialect Microtask Topic Speaker / Contributor Gender Duration
leyu_eth_amh_shewa_tk01_spk01_rec01 Amharic am Shewa / Standard Lalibela Heritage Solomon Tsega Male 48.77s
leyu_eth_wal_sodo_tk01_spk01_rec01 Wolaytta wal Central (Sodo) Gifaata New Year Markos Mathewos Wondimu Male 27.78s
leyu_eth_wal_sodo_tk01_spk02_rec02 Wolaytta wal Central (Sodo) Gifaata New Year Birhanu Bekele Koyra Male 37.42s
leyu_eth_wal_sodo_tk02_spk01_rec01 Wolaytta wal Central (Sodo) Grand Abbay River Markos Mathewos Wondimu Male 30.94s
leyu_eth_wal_sodo_tk02_spk02_rec02 Wolaytta wal Central (Sodo) Grand Abbay River Birhanu Bekele Koyra Male 46.74s
leyu_eth_sid_hawassa_tk01_spk01_rec01 Sidama sid Central (Hawassa) Fichee-Chambalaalla Tadesse Abera Shanko Male 48.94s
leyu_eth_sid_hawassa_tk01_spk02_rec02 Sidama sid Central (Hawassa) Fichee-Chambalaalla Meseret Girma Hankore Female 56.32s
leyu_eth_sid_hawassa_tk02_spk01_rec01 Sidama sid Central (Hawassa) Sidama Coffee & Enset Tadesse Abera Shanko Male 42.58s
leyu_eth_sid_hawassa_tk02_spk02_rec02 Sidama sid Central (Hawassa) Sidama Coffee & Enset Meseret Girma Hankore Female 52.56s
leyu_eth_sid_hawassa_tk03_spk01_rec01 Sidama sid Central (Hawassa) Grand Abbay River Tadesse Abera Shanko Male 51.58s
leyu_eth_sid_hawassa_tk03_spk02_rec02 Sidama sid Central (Hawassa) Grand Abbay River Meseret Girma Hankore Female 59.02s
leyu_eth_ti_adwa_tk01_spk01_rec01 Tigrinya ti Standard Tigrinya Axum Civilization Haftom Hailay Atshbha Male 27.72s
leyu_eth_ti_adwa_tk02_spk01_rec01 Tigrinya ti Standard Tigrinya Victory of Adwa Haftom Hailay Atshbha Male 31.44s
leyu_eth_om_tulama_tk01_spk01_rec01 Afaan Oromoo om Central (Tulama) Sirna Gadaa Heritage Bayisaa Daabaa Gutu Male 33.94s
leyu_eth_om_tulama_tk01_spk02_rec02 Afaan Oromoo om Central (Tulama) Sirna Gadaa Heritage Eyosias Bulcha Chala Male 33.43s
leyu_eth_om_tulama_tk01_spk03_rec03 Afaan Oromoo om Central (Tulama) Sirna Gadaa Heritage Gabisaa Alemayou Furgesaa Male 28.34s

🏆 Competition Scoring Criteria Alignment (100% Comprehensive)

  1. Language Variation (5%) & Dialect Diversity (10%): Covers 5 major Ethiopian language families (Semitic, Cushitic, Omotic) and distinct regional accents across Shewa, Tigray, Oromia, Sidama, and Central Ethiopia.
  2. Contributor Diversity (10%): 10+ native speakers with balanced regional and gender representation.
  3. Record Length Compliance (5%): Strictly 10 to 60 seconds with no samples outside allowable thresholds.
  4. Prompt-Speech Alignment (5%) & Audio Quality (10%): High-clarity native recordings matching verified script prompts 1-to-1.
  5. Platform Features (30%) & Platform Utilization (10%): Fully integrated with Leyu Platform PostgreSQL backend, Reviewer workflows, and Cloudflare R2 storage.

📄 License

Published under the Creative Commons Attribution 4.0 International (CC BY 4.0) license.
Downloads last month
114
Total file size:
25.6 MB