dataset_info.json

{
  "name": "Pashto Synthetic Speech Dataset Parquet (5k)",
  "version": "1.0.0",
  "description": "Synthetic Pashto speech dataset with 5000 sentences in male and female voices (Parquet format)",
  "license": "MIT",
  "created_at": "2025-04-16T23:00:45.737540",
  "format": "parquet",
  "statistics": {
    "total_sentences": 5000,
    "total_male_recordings": 5000,
    "total_female_recordings": 5000,
    "total_recordings": 10000,
    "sampling_rate": 24000,
    "total_male_duration_seconds": 4310.241,
    "total_female_duration_seconds": 4396.275,
    "average_male_duration_seconds": 0.8620482,
    "average_female_duration_seconds": 0.8792549999999999,
    "longest_male_clip_seconds": 2.001,
    "longest_female_clip_seconds": 2.031,
    "shortest_male_clip_seconds": 0.228,
    "shortest_female_clip_seconds": 0.231
  },
  "voices": {
    "male": "ps-AF-GulNawazNeural",
    "female": "ps-AF-LatifaNeural"
  },
  "audio_format": "WAV, 24kHz, 16-bit PCM"
}