{
  "name": "multiwoz_dst",
  "description": "MultiWOZ dataset prepared for dialogue state tracking with aligned audio and dialogue context.",
  "license": "CC0-1.0",
  "license_notes": "This dataset is released under CC0 1.0 Universal (Public Domain).",
  "citation": "https://arxiv.org/pdf/2511.22503",
  "credits": ["https://huggingface.co/datasets/pirxus/spokenwoz-whisper"],
  "features": {
    "text": "string",
    "normalized_text": "string",
    "context": "string",
    "normalized_context": "string",
    "audio": "audio/16000",
    "audio_len": "int32",
    "wav_id": "string",
    "turn_index": "int32",
    "agent_text": "string",
    "domains": "string",
    "slots": "string"
  },
  "splits": {
    "train": {"num_examples": 56750, "num_bytes": 6119340859},
    "validation": {"num_examples": 7374, "num_bytes": 1278573139}
  },
  "version": "0.0.0",
  "notes": "Dataset card updated with CC0 license and paper reference."
}