You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this dataset content.

Voice Synthetic Dataset

A Dhivehi text-to-speech dataset containing professionally recorded audio samples with their corresponding transcriptions.

Dataset Configuration

  • Audio Format: WAV files
  • Language: Dhivehi (dv)
  • Source: cv
  • Split Ratio:
    • Train: 80%
    • Test: 20%
  • Structure:
    • audio: Path to the audio file
    • sentence: Corresponding Dhivehi text transcription
    • source: Source of the audio data

Data Format

{
    'audio': {
        'path': 'path/to/audio.wav',
        'array': array(...),
        'sampling_rate': 22050
    },
    'sentence': 'ދިވެހި ޓެކްސްޓް'  # Dhivehi text
    'source': 'cv'  # Source identifier
}
Downloads last month
12

Models trained or fine-tuned on alakxender/voice-synthetic