README.md

metadata

language:
  - uk
license: apache-2.0
task_categories:
  - text-to-speech
configs:
  - config_name: train
    data_files:
      - split: train
        path: train.parquet

Voice MYKYTA (male)

🇺🇦 Open Source Ukrainian Text-to-Speech dataset named Mykyta

Community

• Discord: https://discord.gg/yVAjkBgmt4
• Speech Recognition: https://t.me/speech_recognition_uk
• Speech Synthesis: https://t.me/speech_synthesis_uk

Overview

More details about this dataset - https://github.com/egorsmkv/ukrainian-tts-datasets/tree/main/mykyta

License: Apache 2.0

Listen to DEMO (choose "mykyta" in the Voice field)

Features

• Quality: high
• Duration: 8h10m
• Audio formats: OPUS/WAV
• Text format: JSONL (a metadata.jsonl file)
• Frequency: 16000/22050/48000 Hz

Original version

In the OPUS format

• 48000 Hz: https://huggingface.co/datasets/Yehor/ukrainian-tts-mykyta/resolve/main/dataset_mykyta_ogg.zip

Trimmed version (removed silence)

Silence is removed by https://github.com/proger/uk#align-text-to-audio-and-trim-silence

In the WAV format

• 48000 Hz: https://huggingface.co/datasets/Yehor/ukrainian-tts-mykyta/resolve/main/dataset_mykyta_trimmed_48khz.zip
• 22050 Hz: https://huggingface.co/datasets/Yehor/ukrainian-tts-mykyta/resolve/main/dataset_mykyta_trimmed_22khz.zip
• 16000 Hz: https://huggingface.co/datasets/Yehor/ukrainian-tts-mykyta/resolve/main/dataset_mykyta_trimmed_16khz.zip

Development

uv venv --python 3.12

source .venv/bin/activate

uv pip install -r requirements.txt
uv pip install -r requirements-dev.txt