README.md

metadata

dataset_info:
  - config_name: jpn
    features:
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: timestamps_start
        sequence: float64
      - name: timestamps_end
        sequence: float64
      - name: speakers
        sequence: string
    splits:
      - name: data
        num_bytes: 2159798942
        num_examples: 120
    download_size: 2119318800
    dataset_size: 2159798942
configs:
  - config_name: jpn
    data_files:
      - split: data
        path: jpn/data-*
license: cc-by-nc-sa-4.0
language:
  - ja
tags:
  - speaker-diarization
  - speaker-segmentation
  - voice-activity-detection
extra_gated_fields:
  Company: text
  Country: country

talkbank/callhome for JA plus

talkbank/callhomeのJAデータをwavに変換して、metaデータをjsonにまとめたデータセット

meta data

以下のようなデータ構造の配列になっています

    {
        "audio_filename": "callhome_jpn_0.wav",
        "sampling_rate": 16000,
        "duration": 606.53,
        "utterances": [
            {
                "start_time": 0.0,
                "end_time": 3.4799999999999898,
                "speaker": "B"
            },
            {
                "start_time": 1.5999999999999943,
                "end_time": 2.059999999999988,
                "speaker": "A"
            }
            ...
        ]
    }

評価用にrttmファイルも追加しています

SPEAKER callhome_jpn_0 1 0.000 3.480 <NA> <NA> B <NA> <NA>
SPEAKER callhome_jpn_0 1 1.600 0.460 <NA> <NA> A <NA> <NA>