{"success":true,"data":{"author":"Junbo Zhang et al. (Xiaomi, SpeechOcean)","bucketAccessType":"public","bucketName":"mispeech/speechocean762","bucketPrefix":"","category":"Speech & Audio","cloudProvider":"huggingface","createdAt":"2026-11-14T14:15:17.566Z","description":"speechocean762 is an open corpus for automatic pronunciation assessment in computer-assisted language learning. It consists of 5,000 English sentences read by non-native speakers whose mother tongue is Mandarin; half of the speakers are children and half adults, with age and gender provided. Five human experts scored every utterance independently under the same rubric, giving manual annotations at sentence level (accuracy, completeness, fluency, prosody, total), word level (accuracy, stress, total) and phoneme level (per-phone accuracy and marked mispronunciations). It is the standard benchmark for goodness-of-pronunciation (GOP) and pronunciation scoring models and for mispronunciation detection and diagnosis. This repository holds the corpus as two Parquet files under data/ (train and test, 2,500 utterances each per the card example), with the audio and nested word and phone annotations (measured 2026-11-14; the Hub reports about 2 GB). Load with datasets.load_dataset('mispeech/speechocean762', split='test') or pandas. Released on OpenSLR (SLR101) under CC BY 4.0 for commercial and non-commercial use; cite Zhang et al., Interspeech 2021.","displayTitle":"speechocean762: non-native English pronunciation scoring corpus","id":"7a1e6e49-4d5f-54b6-a9b1-920eb280e028","license":"CC-BY-4.0","licenseEvidenceUrl":"https://www.openslr.org/101/","objectCountEstimate":4,"probeStatus":"ok","referenceUrl":"https://arxiv.org/abs/2104.01378","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"mispeech/speechocean762\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"mispeech/speechocean762\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=mispeech, license:apache-2.0, size_categories:1K<n<10K, format:parquet, downloads=3324"],"format":"mixed","generatedAt":"2026-11-14T14:15:20.708Z","layout":{"example":"data/test-00000-of-00001.parquet","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","encoding":"hex","key":"data/test-00000-of-00001.parquet","truncatedAt":512}],"sampleObjects":["data/test-00000-of-00001.parquet","data/train-00000-of-00001.parquet"],"sidecars":[{"bytes":"---\nlanguage:\n- en\nlicense: apache-2.0\nsize_categories:\n- 1K<n<10K\ntask_categories:\n- automatic-speech-recognition\npretty_name: speechocean762\ntags:\n- pronunciation-scoring\n- arxiv:2104.01378\nconfigs:\n- config_name: default\n  data_files:\n  - split: train\n    path: data/train-*\n  - split: test\n    path: data/test-*\ndataset_info:\n  features:\n  - name: accuracy\n    dtype: int64\n  - name: completeness\n    dtype: float64\n  - name: fluency\n    dtype: int64\n  - name: prosodic\n    dtype: int64\n  - name: text\n    dtype: string\n  - name: total\n    dtype: int64\n  - name: words\n    list:\n    - name: accuracy\n      dtype: int64\n    - name: phones\n      sequence: string\n    - name: phones-accuracy\n      sequence: float64\n    - name: stress\n      dtype: int64\n    - name: text\n      dtype: string\n    - name: total\n      dtype: int64\n    - name: mispronunciations\n      list:\n      - name: canonical-phone\n        dtype: string\n      - name: index\n        dtype: int64\n      - name: pronounced-phone\n        dtype: string\n  - name: speaker\n    dtype: string\n  - name: gender\n    dtype: string\n  - name: age\n    dtype: int64\n  - name: audio\n    dtype: audio\n  splits:\n  - name: train\n    num_bytes: 333140374\n    num_examples: 2500\n  - name: test\n    num_bytes: 311843439\n    num_examples: 2500\n  download_size: 644387157\n  dataset_size: 644983813\n---\n# speechocean762: A non-native English corpus for pronunciation scoring task\n\n## Introduction\nPronunciation scoring is a crucial technology in computer-assisted language learning (CALL) systems. The pronunciation quality scores might be given at phoneme-level, word-level, and sentence-level for a typical pronunciation scoring task.\n\nThis corpus aims to provide a free public dataset for the pronunciation scoring task.\nKey features:\n* It is available for free download for both commercial and non-commercial purposes.\n* The speaker variety encompasses young children and adults.\n* The manual annotations are in multiple aspects at sentence-level, word-level and phoneme-level.\n\nThis corpus consists of 5000 English sentences. All the speakers are non-native, and their mother tongue is Mandarin. Half of the speakers are Children, and the others are adults. The information of age and gender are provided.\n\nFive experts made the scores. To avoid subjective bias, each expert scores independently under the same metric.\n\n## Uses\n```python\n>>> from datasets import load_dataset\n\n>>> test_set = load_dataset(\"mispeech/speechocean762\", split=\"test\")\n\n>>> len(test_set)\n2500\n\n>>> next(iter(test_set))\n{'accuracy': 9,\n 'completeness': 10.0,\n 'fluency': 9,\n 'prosodic': 9,\n 'text': 'MARK IS GOING TO SEE ELEPHANT',\n 'total': 9,\n 'words': [{'accuracy': 10,\n   'phones': ['M', 'AA0', 'R', 'K'],\n   'phones-accuracy': [2.0, 2.0, 1.8, 2.0],\n   'stress': 10,\n   'text': 'MARK',\n   'total': 10,\n   'mispronunciations': []},\n  {'accuracy': 10,\n   'phones': ['IH0', 'Z'],\n   'phones-accuracy': [2.0, 1.8],\n   'stress': 10,\n   'text': 'IS',\n   'total': 10,\n   'mispronunciations': []},\n  {'accuracy': 10,\n   'phones': ['G', 'OW0', 'IH0', 'NG'],\n   'phones-accuracy': [2.0, 2.0, 2.0, 2.0],\n   'stress': 10,\n   'text': 'GOING',\n   'total': 10,\n   'mispronunciations': []},\n  {'accuracy': 10,\n   'phones': ['T', 'UW0'],\n   'phones-accuracy': [2.0, 2.0],\n   'stress': 10,\n   'text': 'TO',\n   'total': 10,\n   'mispronunciations': []},\n  {'accuracy': 10,\n   'phones': ['S', 'IY0'],\n   'phones-accuracy': [2.0, 2.0],\n   'stress': 10,\n   'text': 'SEE',\n   'total': 10,\n   'mispronunciations': []},\n  {'accuracy': 10,\n   'phones': ['EH1', 'L', 'IH0', 'F', 'AH0', 'N', 'T'],\n   'phones-accuracy': [2.0, 2.0, 2.0, 2.0, 2.0, 2.0, 2.0],\n   'stress': 10,\n   'text': 'ELEPHANT',\n   'total': 10,\n   'mispronunciations': []}],\n 'speaker': '0003',\n 'gender': 'm',\n 'age': 6,\n 'audio': {'path': '000030012.wav',\n  'array': array([-0.00119019, -0.00500488, -0.00283813, ...,  0.00274658,\n          0.        ,  0.00125122]),\n  'sampling_rate': 16000}}\n```\n\n## The scoring metric\nThe experts score at three levels: phoneme-level, word-level, and sentence-level.\n\n### Sentence level\nSco","encoding":"utf8","key":"README.md","name":"README.md","size":7548,"truncated":true}],"sourceObjectKey":"data/test-00000-of-00001.parquet","starterCell":"from datasets import load_dataset\nds = load_dataset(\"mispeech/speechocean762\")\nprint(ds)","starterSamplePath":"data/test-00000-of-00001.parquet","totalObjects":4},"summary":"speechocean762 is an open corpus for automatic pronunciation assessment in computer-assisted language learning.","title":"speechocean762: non-native English pronunciation scoring corpus","updatedAt":"2026-11-14T17:38:04.926Z","visibility":"public","websiteUrl":"https://www.openslr.org/101/"}}