{"success":true,"data":{"author":"MLCommons (Daniel Galvez et al.)","bucketAccessType":"public","bucketName":"MLCommons/peoples_speech","bucketPrefix":"","category":"Speech & Audio","cloudProvider":"huggingface","createdAt":"2026-11-14T14:15:17.612Z","description":"The People's Speech is among the largest English speech recognition corpora, with more than 30,000 hours of transcribed English speech from a diverse set of speakers, collected by MLCommons from openly licensed audio on the Internet Archive, including legal and government proceedings, oral histories and speeches. It was built to give speech-to-text research and industry a corpus large enough to train modern ASR systems under a permissive licence. Data are split by licence and by transcript quality: cc-by-clean, cc-by-dirty, cc-by-sa-clean, cc-by-sa-dirty, plus a small microset and shared validation and test sets. This repository holds those as seven folders (clean, clean_sa, dirty, dirty_sa, microset, test, validation) of Parquet shards, for example 818 files in clean (measured 2026-11-14; the Hub reports about 4.8 TB). Each row holds an id, 16 kHz FLAC audio, duration in milliseconds and the transcript. Load with datasets.load_dataset('MLCommons/peoples_speech', 'clean', streaming=True) or read shards with pyarrow. This is the supervised corpus, distinct from the Unsupervised People's Speech release. MLCommons licenses the subsets under CC-BY and CC-BY-SA 4.0 for academic and commercial use; keep the subset's licence when redistributing.","displayTitle":"The People's Speech: 30,000-hour English speech recognition corpus","id":"7894b47e-8924-5292-b901-84da6b644482","license":"CC-BY-SA-4.0","licenseEvidenceUrl":"https://huggingface.co/datasets/MLCommons/peoples_speech","objectCountEstimate":4498,"probeStatus":"ok","referenceUrl":"https://arxiv.org/abs/2111.09344","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"MLCommons/peoples_speech\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"MLCommons/peoples_speech\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=MLCommons, license:cc-by-2.0, size_categories:1M<n<10M, format:parquet, downloads=46790"],"format":"mixed","generatedAt":"2026-11-14T14:15:22.859Z","layout":{"example":"clean/test-00000-of-00009.parquet","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","encoding":"hex","key":"clean/test-00000-of-00009.parquet","truncatedAt":512}],"sampleObjects":["clean/test-00000-of-00009.parquet","clean/test-00001-of-00009.parquet","clean/test-00002-of-00009.parquet","clean/test-00003-of-00009.parquet","clean/test-00004-of-00009.parquet"],"sidecars":[{"bytes":"---\nannotations_creators:\n- crowdsourced\n- machine-generated\nlanguage_creators:\n- crowdsourced\n- machine-generated\nlanguage:\n- en\nlicense:\n- cc-by-2.0\n- cc-by-2.5\n- cc-by-3.0\n- cc-by-4.0\n- cc-by-sa-3.0\n- cc-by-sa-4.0\nmultilinguality:\n- monolingual\nsize_categories:\n- 1T<n\nsource_datasets:\n- original\ntask_categories:\n- automatic-speech-recognition\ntask_ids: []\npretty_name: People's Speech\ntags:\n- robust-speech-recognition\n- noisy-speech-recognition\n- speech-recognition\ndataset_info:\n- config_name: clean\n  features:\n  - name: id\n    dtype: string\n  - name: audio\n    dtype:\n      audio:\n        sampling_rate: 16000\n  - name: duration_ms\n    dtype: int32\n  - name: text\n    dtype: string\n  splits:\n  - name: train\n    num_bytes: 401733771186.124\n    num_examples: 1501271\n  - name: validation\n    num_bytes: 2459781412.24\n    num_examples: 18622\n  - name: test\n    num_bytes: 4324307722.96\n    num_examples: 34898\n  download_size: 398550700437\n  dataset_size: 408517860321.32404\n- config_name: clean_sa\n  features:\n  - name: id\n    dtype: string\n  - name: audio\n    dtype:\n      audio:\n        sampling_rate: 16000\n  - name: duration_ms\n    dtype: int32\n  - name: text\n    dtype: string\n  splits:\n  - name: train\n    num_bytes: 75267509124.558\n    num_examples: 257093\n  - name: validation\n    num_bytes: 2075929254.254\n    num_examples: 18622\n  - name: test\n    num_bytes: 3894954757.41\n    num_examples: 34898\n  download_size: 72518549222\n  dataset_size: 81238393136.222\n- config_name: dirty\n  features:\n  - name: id\n    dtype: string\n  - name: audio\n    dtype:\n      audio:\n        sampling_rate: 16000\n  - name: duration_ms\n    dtype: int32\n  - name: text\n    dtype: string\n  splits:\n  - name: train\n    num_bytes: 1569500875399.994\n    num_examples: 5476898\n  - name: validation\n    num_bytes: 2641406179.2539997\n    num_examples: 18622\n  - name: test\n    num_bytes: 5097236056.41\n    num_examples: 34898\n  download_size: 1496747948260\n  dataset_size: 1577239517635.6577\n- config_name: dirty_sa\n  features:\n  - name: id\n    dtype: string\n  - name: audio\n    dtype:\n      audio:\n        sampling_rate: 16000\n  - name: duration_ms\n    dtype: int32\n  - name: text\n    dtype: string\n  splits:\n  - name: train\n    num_bytes: 163776914241.91\n    num_examples: 548014\n  - name: validation\n    num_bytes: 2075929254.254\n    num_examples: 18622\n  - name: test\n    num_bytes: 3894954757.41\n    num_examples: 34898\n  download_size: 149326092074\n  dataset_size: 169747798253.574\n- config_name: microset\n  features:\n  - name: id\n    dtype: string\n  - name: audio\n    dtype:\n      audio:\n        sampling_rate: 16000\n  - name: duration_ms\n    dtype: int32\n  - name: text\n    dtype: string\n  splits:\n  - name: train\n    num_bytes: 92397066.0\n    num_examples: 336\n  download_size: 90204303\n  dataset_size: 92397066.0\n- config_name: test\n  features:\n  - name: id\n    dtype: string\n  - name: audio\n    dtype:\n      audio:\n        sampling_rate: 16000\n  - name: duration_ms\n    dtype: int32\n  - name: text\n    dtype: string\n  splits:\n  - name: test\n    num_bytes: 3894954757.41\n    num_examples: 34898\n  download_size: 4087772459\n  dataset_size: 3894954757.41\n- config_name: validation\n  features:\n  - name: id\n    dtype: string\n  - name: audio\n    dtype:\n      audio:\n        sampling_rate: 16000\n  - name: duration_ms\n    dtype: int32\n  - name: text\n    dtype: string\n  splits:\n  - name: validation\n    num_bytes: 2075929254.254\n    num_examples: 18622\n  download_size: 2335244149\n  dataset_size: 2075929254.254\nconfigs:\n- config_name: clean\n  data_files:\n  - split: train\n    path: clean/train-*\n  - split: validation\n    path: clean/validation-*\n  - split: test\n    path: clean/test-*\n- config_name: clean_sa\n  data_files:\n  - split: train\n    path: clean_sa/train-*\n  - split: validation\n    path: clean_sa/validation-*\n  - split: test\n    path: clean_sa/test-*\n- config_name: dirty\n  data_files:\n  - split: train\n    path: dirty/train-*\n  - split: validation\n    path: dirty/validation-*\n  - split: test\n    path: dirty/test-*\n- config_name: dirty_sa\n  data_files:\n  - split: train\n    path: dirty_sa","encoding":"utf8","key":"README.md","name":"README.md","size":11530,"truncated":true}],"sourceObjectKey":"clean/test-00000-of-00009.parquet","starterCell":"from datasets import load_dataset\nds = load_dataset(\"MLCommons/peoples_speech\")\nprint(ds)","starterSamplePath":"clean/test-00000-of-00009.parquet","totalObjects":4498},"summary":"The People's Speech is among the largest English speech recognition corpora, with more than 30,000 hours of transcribed English speech from a diverse set of speakers, collected by…","title":"The People's Speech: 30,000-hour English speech recognition corpus","updatedAt":"2026-11-14T17:38:15.988Z","visibility":"public","websiteUrl":"https://mlcommons.org/en/peoples-speech/"}}