{
  "archive": "dataset.tar.zst",
  "archive_sha256": "3636eb106f015593dfbde4cf7bd852235bcf757b188c3a401c459e71408dfe8a",
  "archive_size_bytes": 45798638216,
  "build_config_fingerprint": "82ce51182bdb5e0a52873a5ba90f6c6a854f8d5212c8631c5fd13197e0c3d59c",
  "checksum": "dataset.tar.zst.sha256",
  "corpus_ids": [
    "ears",
    "hifi_tts",
    "vctk"
  ],
  "created_at": "2026-08-23T19:27:53.553407+00:00",
  "dataset_fingerprint": "1a0cb5629d0a7f90ab673aed245e7f13e7457c461d82dea96694001a0f58b595",
  "dataset_id": "speech-hard",
  "dataset_schema_version": "declip_dataset_v8",
  "dataset_version": "1",
  "release_tag": "speech-hard-v1",
  "repository": "anthonyleedom/test2",
  "schema_version": "declip_dataset_release_v1",
  "segment_samples": 288000,
  "split_counts": {
    "test": 10468,
    "train": 82443,
    "validation": 10485
  }
}