Upload README.md with huggingface_hub

README.md

@@ -1,484 +1,82 @@
 ---
-dataset_info:
-- config_name: ar-SA
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 1397432
-    num_examples: 8
-  download_size: 964431
-  dataset_size: 1397432
-- config_name: ca-ES
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 1397432
-    num_examples: 8
-  download_size: 3469988
-  dataset_size: 1397432
-- config_name: cs-CZ
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 1397432
-    num_examples: 8
-  download_size: 1332607
-  dataset_size: 1397432
-- config_name: de-DE
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 2794864
-    num_examples: 16
-  download_size: 2091323
-  dataset_size: 2794864
-- config_name: en-AU
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 1397432
-    num_examples: 8
-  download_size: 976846
-  dataset_size: 1397432
-- config_name: en-GB
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 1397432
-    num_examples: 8
-  download_size: 1077943
-  dataset_size: 1397432
-- config_name: en-US
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 1397432
-    num_examples: 8
-  download_size: 1658520
-  dataset_size: 1397432
-- config_name: es-ES
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 2794864
-    num_examples: 16
-  download_size: 2179941
-  dataset_size: 2794864
-- config_name: fr-FR
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 2794864
-    num_examples: 16
-  download_size: 1222592
-  dataset_size: 2794864
-- config_name: gl-ES
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 1397432
-    num_examples: 8
-  download_size: 2503532
-  dataset_size: 1397432
-- config_name: hu-HU
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 1397432
-    num_examples: 8
-  download_size: 1110414
-  dataset_size: 1397432
-- config_name: it-IT
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 1397432
-    num_examples: 8
-  download_size: 2361540
-  dataset_size: 1397432
-- config_name: ko-KR
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 2794864
-    num_examples: 16
-  download_size: 1362853
-  dataset_size: 2794864
-- config_name: nl-NL
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 2794864
-    num_examples: 16
-  download_size: 2751170
-  dataset_size: 2794864
-- config_name: pl-PL
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 2794864
-    num_examples: 16
-  download_size: 3963790
-  dataset_size: 2794864
-- config_name: pt-PT
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 2794864
-    num_examples: 16
-  download_size: 3343754
-  dataset_size: 2794864
-- config_name: ru-RU
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 2794864
-    num_examples: 16
-  download_size: 3078725
-  dataset_size: 2794864
-- config_name: tr-TR
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 1397432
-    num_examples: 8
-  download_size: 983589
-  dataset_size: 1397432
-- config_name: vi-VN
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 1397432
-    num_examples: 8
-  download_size: 1174096
-  dataset_size: 1397432
-- config_name: zh-CN
-  features:
-  - name: sample_id
-    dtype: string
-  - name: audio
-    dtype: audio
-  - name: transcript
-    dtype: string
-  - name: lang
-    dtype: string
-  - name: domain
-    dtype: string
-  - name: source_dataset
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 1397432
-    num_examples: 8
-  download_size: 1575801
-  dataset_size: 1397432
-configs:
-- config_name: ar-SA
-  data_files:
-  - split: train
-    path: ar-SA/train-*
-- config_name: ca-ES
-  data_files:
-  - split: train
-    path: ca-ES/train-*
-- config_name: cs-CZ
-  data_files:
-  - split: train
-    path: cs-CZ/train-*
-- config_name: de-DE
-  data_files:
-  - split: train
-    path: de-DE/train-*
-- config_name: en-AU
-  data_files:
-  - split: train
-    path: en-AU/train-*
-- config_name: en-GB
-  data_files:
-  - split: train
-    path: en-GB/train-*
-- config_name: en-US
-  data_files:
-  - split: train
-    path: en-US/train-*
-- config_name: es-ES
-  data_files:
-  - split: train
-    path: es-ES/train-*
-- config_name: fr-FR
-  data_files:
-  - split: train
-    path: fr-FR/train-*
-- config_name: gl-ES
-  data_files:
-  - split: train
-    path: gl-ES/train-*
-- config_name: hu-HU
-  data_files:
-  - split: train
-    path: hu-HU/train-*
-- config_name: it-IT
-  data_files:
-  - split: train
-    path: it-IT/train-*
-- config_name: ko-KR
-  data_files:
-  - split: train
-    path: ko-KR/train-*
-- config_name: nl-NL
-  data_files:
-  - split: train
-    path: nl-NL/train-*
-- config_name: pl-PL
-  data_files:
-  - split: train
-    path: pl-PL/train-*
-- config_name: pt-PT
-  data_files:
-  - split: train
-    path: pt-PT/train-*
-- config_name: ru-RU
-  data_files:
-  - split: train
-    path: ru-RU/train-*
-- config_name: tr-TR
-  data_files:
-  - split: train
-    path: tr-TR/train-*
-- config_name: vi-VN
-  data_files:
-  - split: train
-    path: vi-VN/train-*
-- config_name: zh-CN
-  data_files:
-  - split: train
-    path: zh-CN/train-*
+license: cc-by-4.0
+language:
+- multilingual
+task_categories:
+- automatic-speech-recognition
+pretty_name: STT Sampler v1
 ---
+
+# stt-sampler-v1
+
+A small, balanced, deterministic multi-lingual STT eval sample for
+[OVOS Plugin Arena](https://github.com/OpenVoiceOS/ovos-plugin-arena)
+leaderboards. It gives every registered STT locale immediate breadth-first
+coverage — 8 clips per (source, language) pair — instead of waiting for a
+full run over the much larger per-locale corpora the arena also benchmarks
+against (MInDS-14, Speech-MASSIVE, FLEURS).
+
+Clips are selected deterministically: each source split is sorted by its
+own sample id, then strided (`index = i * floor(n / 8)` for i in 0..7) so
+the 8 picks spread across speakers/sessions instead of clustering on the
+corpus's first few rows (which tend to share one speaker). No randomness
+is used. Unavailable/corrupt rows are skipped by advancing to the next
+candidate in sorted order, deterministically.
+
+Total clips: **224**
+
+## Sources and licenses
+
+| Source | HF id | License | Domain |
+|---|---|---|---|
+| MInDS-14 | [PolyAI/minds14](https://huggingface.co/datasets/PolyAI/minds14) | CC-BY-4.0 | banking (telephone intent calls) |
+| Speech-MASSIVE | [FBK-MT/Speech-MASSIVE-test](https://huggingface.co/datasets/FBK-MT/Speech-MASSIVE-test) | CC-BY-NC-SA-4.0 | assistant (spoken voice-assistant utterances) |
+| FLEURS | [google/fleurs](https://huggingface.co/datasets/google/fleurs) (ca_es, gl_es) | CC-BY-4.0 | read-speech |
+
+## Columns
+
+| column | type | description |
+|---|---|---|
+| `sample_id` | str | `<source>/<orig_id>` — source-tagged, traceable back to the origin row |
+| `audio` | `Audio` | embedded audio, kept at the source's native sampling rate |
+| `transcript` | str | reference transcript |
+| `lang` | str | BCP-47 locale tag |
+| `domain` | str | `banking` / `assistant` / `read-speech` |
+| `source_dataset` | str | origin HF dataset id |
+
+Stored as parquet with embedded audio, split by `lang` config so the arena
+runner can resolve one HF split per language.
+
+## Per-(source, language) counts
+
+| source_dataset | lang | domain | n |
+|---|---|---|---|
+| FBK-MT/Speech-MASSIVE-test | ar-SA | assistant | 8 |
+| FBK-MT/Speech-MASSIVE-test | de-DE | assistant | 8 |
+| FBK-MT/Speech-MASSIVE-test | es-ES | assistant | 8 |
+| FBK-MT/Speech-MASSIVE-test | fr-FR | assistant | 8 |
+| FBK-MT/Speech-MASSIVE-test | hu-HU | assistant | 8 |
+| FBK-MT/Speech-MASSIVE-test | ko-KR | assistant | 8 |
+| FBK-MT/Speech-MASSIVE-test | nl-NL | assistant | 8 |
+| FBK-MT/Speech-MASSIVE-test | pl-PL | assistant | 8 |
+| FBK-MT/Speech-MASSIVE-test | pt-PT | assistant | 8 |
+| FBK-MT/Speech-MASSIVE-test | ru-RU | assistant | 8 |
+| FBK-MT/Speech-MASSIVE-test | tr-TR | assistant | 8 |
+| FBK-MT/Speech-MASSIVE-test | vi-VN | assistant | 8 |
+| PolyAI/minds14 | cs-CZ | banking | 8 |
+| PolyAI/minds14 | de-DE | banking | 8 |
+| PolyAI/minds14 | en-AU | banking | 8 |
+| PolyAI/minds14 | en-GB | banking | 8 |
+| PolyAI/minds14 | en-US | banking | 8 |
+| PolyAI/minds14 | es-ES | banking | 8 |
+| PolyAI/minds14 | fr-FR | banking | 8 |
+| PolyAI/minds14 | it-IT | banking | 8 |
+| PolyAI/minds14 | ko-KR | banking | 8 |
+| PolyAI/minds14 | nl-NL | banking | 8 |
+| PolyAI/minds14 | pl-PL | banking | 8 |
+| PolyAI/minds14 | pt-PT | banking | 8 |
+| PolyAI/minds14 | ru-RU | banking | 8 |
+| PolyAI/minds14 | zh-CN | banking | 8 |
+| google/fleurs | ca-ES | read-speech | 8 |
+| google/fleurs | gl-ES | read-speech | 8 |
+