{"success":true,"data":{"author":"Yuma Koizumi, Heiga Zen, Shigeki Karita et al. (Google)","bucketAccessType":"public","bucketName":"mythicinfinity/libritts_r","bucketPrefix":"","category":"Speech & Audio","cloudProvider":"huggingface","createdAt":"2026-11-14T14:15:17.357Z","description":"LibriTTS-R is a sound-quality-improved version of the LibriTTS corpus, the roughly 585-hour, 24 kHz multi-speaker English read-speech corpus built for text-to-speech research. The utterances are identical to LibriTTS; only the audio was changed, by running every sample through Miipher, a speech restoration model from Google that removes noise, reverberation and codec artefacts. The authors show that TTS models trained on LibriTTS-R produce more natural speech than the same models trained on LibriTTS, which has made it a common training set for zero-shot TTS, speech language models and vocoders. This repository is a Parquet conversion of OpenSLR SLR141: a data folder with seven split folders (dev.clean, dev.other, test.clean, test.other, train.clean.100, train.clean.360, train.other.500); train.other.500 alone holds 103 Parquet shards of about 53 GB (measured 2026-11-14; the Hub reports about 101 GB for the repo). Columns include audio, normalised and original text, speaker id, chapter id and utterance id. Load with datasets.load_dataset('mythicinfinity/libritts_r', 'clean') or read shards with pyarrow. Licensed CC BY 4.0 by the publishers on OpenSLR; cite Koizumi et al., Interspeech 2023.","displayTitle":"LibriTTS-R: restored LibriTTS speech corpus","id":"b1de45fc-b281-52b0-9d2a-d495c4338f51","license":"CC-BY-4.0","licenseEvidenceUrl":"https://www.openslr.org/141/","objectCountEstimate":198,"probeStatus":"ok","referenceUrl":"https://arxiv.org/abs/2305.18802","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"mythicinfinity/libritts_r\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"mythicinfinity/libritts_r\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=mythicinfinity, license:cc-by-4.0, size_categories:100K<n<1M, format:parquet, downloads=16620"],"format":"mixed","generatedAt":"2026-11-14T14:15:28.109Z","layout":{"example":"data/dev.clean/dev.clean-00000-of-00004.parquet","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","encoding":"hex","key":"data/dev.clean/dev.clean-00000-of-00004.parquet","truncatedAt":512}],"sampleObjects":["data/dev.clean/dev.clean-00000-of-00004.parquet","data/dev.clean/dev.clean-00001-of-00004.parquet","data/dev.clean/dev.clean-00002-of-00004.parquet","data/dev.clean/dev.clean-00003-of-00004.parquet","data/dev.other/dev.other-00000-of-00002.parquet"],"sidecars":[{"bytes":"---\nlicense: cc-by-4.0\ntask_categories:\n- text-to-speech\nlanguage:\n- en\nsize_categories:\n- 10K<n<100K\n\nconfigs:\n- config_name: dev\n  data_files:\n  - split: dev.clean\n    path: \"data/dev.clean/dev.clean*.parquet\"\n    \n- config_name: clean\n  data_files:\n  - split: dev.clean\n    path: \"data/dev.clean/dev.clean*.parquet\"\n  - split: test.clean\n    path: \"data/test.clean/test.clean*.parquet\"\n  - split: train.clean.100\n    path: \"data/train.clean.100/train.clean.100*.parquet\"\n  - split: train.clean.360\n    path: \"data/train.clean.360/train.clean.360*.parquet\"\n    \n- config_name: other\n  data_files:\n  - split: dev.other\n    path: \"data/dev.other/dev.other*.parquet\"\n  - split: test.other\n    path: \"data/test.other/test.other*.parquet\"\n  - split: train.other.500\n    path: \"data/train.other.500/train.other.500*.parquet\"\n    \n- config_name: all\n  data_files:\n  - split: dev.clean\n    path: \"data/dev.clean/dev.clean*.parquet\"\n  - split: dev.other\n    path: \"data/dev.other/dev.other*.parquet\"\n  - split: test.clean\n    path: \"data/test.clean/test.clean*.parquet\"\n  - split: test.other\n    path: \"data/test.other/test.other*.parquet\"\n  - split: train.clean.100\n    path: \"data/train.clean.100/train.clean.100*.parquet\"\n  - split: train.clean.360\n    path: \"data/train.clean.360/train.clean.360*.parquet\"\n  - split: train.other.500\n    path: \"data/train.other.500/train.other.500*.parquet\"\n---\n# Dataset Card for LibriTTS-R\n\n<!-- Provide a quick summary of the dataset. -->\n\nLibriTTS-R [1] is a sound quality improved version of the LibriTTS corpus \n(http://www.openslr.org/60/) which is a multi-speaker English corpus of approximately \n585 hours of read English speech at 24kHz sampling rate, published in 2019.\n\n## Overview\n\nThis is the LibriTTS-R dataset, adapted for the `datasets` library.\n\n## Usage\n\n### Splits\n\nThere are 7 splits (dots replace dashes from the original dataset, to comply with hf naming requirements):\n\n- dev.clean\n- dev.other\n- test.clean\n- test.other\n- train.clean.100\n- train.clean.360\n- train.other.500\n\n### Configurations\n\nThere are 3 configurations, each which limits the splits the `load_dataset()` function will download.\n\nThe default configuration is \"all\".\n\n- \"dev\": only the \"dev.clean\" split (good for testing the dataset quickly)\n- \"clean\": contains only \"clean\" splits\n- \"other\": contains only \"other\" splits\n- \"all\": contains only \"all\" splits\n\n### Example\n\nLoading the `clean` config with only the `train.clean.360` split.\n```\nload_dataset(\"blabble-io/libritts_r\", \"clean\", split=\"train.clean.100\")\n```\n\nStreaming is also supported.\n```\nload_dataset(\"blabble-io/libritts_r\", streaming=True)\n```\n\n### Columns\n\n```\n{\n    \"audio\": datasets.Audio(sampling_rate=24_000),\n    \"text_normalized\": datasets.Value(\"string\"),\n    \"text_original\": datasets.Value(\"string\"),\n    \"speaker_id\": datasets.Value(\"string\"),\n    \"path\": datasets.Value(\"string\"),\n    \"chapter_id\": datasets.Value(\"string\"),\n    \"id\": datasets.Value(\"string\"),\n}\n```\n\n### Example Row\n\n```\n{\n  'audio': {\n    'path': '/home/user/.cache/huggingface/datasets/downloads/extracted/5551a515e85b9e463062524539c2e1cb52ba32affe128dffd866db0205248bdd/LibriTTS_R/dev-clean/3081/166546/3081_166546_000028_000002.wav', \n    'array': ..., \n    'sampling_rate': 24000\n  }, \n  'text_normalized': 'How quickly he disappeared!\"',\n  'text_original': 'How quickly he disappeared!\"',\n  'speaker_id': '3081', \n  'path': '/home/user/.cache/huggingface/datasets/downloads/extracted/5551a515e85b9e463062524539c2e1cb52ba32affe128dffd866db0205248bdd/LibriTTS_R/dev-clean/3081/166546/3081_166546_000028_000002.wav', \n  'chapter_id': '166546', \n  'id': '3081_166546_000028_000002'\n}\n```\n\n## Dataset Details\n\n### Dataset Description\n\n- **License:** CC BY 4.0\n\n### Dataset Sources [optional]\n\n<!-- Provide the basic links for the dataset. -->\n\n- **Homepage:** https://www.openslr.org/141/\n- **Paper:** https://arxiv.org/abs/2305.18802\n\n## Citation\n\n<!-- If there is a paper or blog post introducing the dataset, the APA and Bibtex information for that should go in this section. -->\n\n```\n@ARTICLE{Koizumi2023-hs,\n  title       ","encoding":"utf8","key":"README.md","name":"README.md","size":5650,"truncated":true}],"sourceObjectKey":"data/dev.clean/dev.clean-00000-of-00004.parquet","starterCell":"from datasets import load_dataset\nds = load_dataset(\"mythicinfinity/libritts_r\")\nprint(ds)","starterSamplePath":"data/dev.clean/dev.clean-00000-of-00004.parquet","totalObjects":198},"summary":"LibriTTS-R is a sound-quality-improved version of the LibriTTS corpus, the roughly 585-hour, 24 kHz multi-speaker English read-speech corpus built for text-to-speech research.","title":"LibriTTS-R: restored LibriTTS speech corpus","updatedAt":"2026-11-14T17:33:08.472Z","visibility":"public","websiteUrl":"https://www.openslr.org/141/"}}