CoolFace
Datasetpublic

no7z/hsk-sentences-audio

HSK Sentences Audio 4,354 Chinese sentences graded against the official HSK 3.0 levels 1–6, with pinyin, English translations, per-word glosses, grammar tags, and normal/slow synthetic speech. The complete export contains 8,708 MP3 files. Dataset structure The Viewer reads native Parquet from data/train.parquet, avoiding a dependency on Hugging Face's JSON-to-Parquet conversion service. The same 4,354 records are also available as validated JSON Lines in… See the full description on the dataset page: https://huggingface.co/datasets/no7z/hsk-sentences-audio.

sourceHugging Facecc-by-sa-4.0updated 2mo agoView on Hugging Face
0likes648downloads
Dataset Card

HSK Sentences Audio

4,354 Chinese sentences graded against the official HSK 3.0 levels 1–6, with pinyin, English translations, per-word glosses, grammar tags, and normal/slow synthetic speech. The complete export contains 8,708 MP3 files.

Dataset structure

The Viewer reads native Parquet from data/train.parquet, avoiding a dependency on Hugging Face's JSON-to-Parquet conversion service. The same 4,354 records are also available as validated JSON Lines in data/train.jsonl. Important fields:

  • —id, hsk_level, topic, sentence_type
  • —chinese, traditional, pinyin, pinyin_numbered
  • —translation.en, token-level word / pinyin / gloss_en
  • —grammar_points, official grammar_tags
  • —audio.normal, audio.slow, and synthesis metadata
  • —convenience fields audio_normal and audio_slow for dataset viewers

Sentence counts by level: HSK 1: 281; HSK 2: 538; HSK 3: 727; HSK 4: 801; HSK 5: 965; HSK 6: 1,042.

Intended uses and limits

Suitable for flashcards, shadowing, listening practice, SRS, pronunciation interfaces, and Chinese-learning research prototypes. Audio is synthetic (CosyVoice2-0.5B, zh-female-studio) and should be disclosed as such in downstream products. The dataset is educational material, not an assessment instrument or a substitute for native-speaker review.

License and attribution

Dataset: CC-BY-SA-4.0. Glosses and some pinyin data derive from CC-CEDICT (CC-BY-SA); audio was synthesized with Apache-2.0 CosyVoice2. See ATTRIBUTION.md in this dataset export for detailed provenance.