README.md

metadata

language:
  - en
license: mit
task_categories:
  - text-to-speech
  - audio-to-audio
pretty_name: Jenny TTS with Mimi Codes
tags:
  - audio
  - speech
  - mimi
  - codec
  - tts
  - jenny

Jenny TTS with Mimi Codes

This dataset adds Mimi codec codes to reach-vb/jenny_tts_dataset.

Dataset Description

Each sample contains:

• audio: Original Jenny TTS audio resampled to 24kHz (Mimi's native rate)
• codes: 8-layer Mimi codec codes (list of 8 lists of integers)
• transcription: Original text transcription
• transcription_normalised: Normalized transcription

Stats

• Samples: 20,978
• Audio Sample Rate: 24kHz (resampled from original 48kHz)
• Codec: Mimi (kyutai/mimi) with 8 codebooks

Usage

from datasets import load_dataset

ds = load_dataset("mazesmazes/jenny-mimi", split="train")

# Access audio and codes together
sample = ds[0]
audio = sample["audio"]  # {'array': [...], 'sampling_rate': 24000}
codes = sample["codes"]  # 8 lists of codec indices
text = sample["transcription_normalised"]

# Decode codes back to audio (requires moshi_mlx or transformers)
import torch
from transformers import MimiModel

mimi = MimiModel.from_pretrained("kyutai/mimi")
codes_tensor = torch.tensor(codes).unsqueeze(0)  # (1, 8, seq_len)
with torch.no_grad():
    decoded = mimi.decode(codes_tensor)
    waveform = decoded.audio_values  # (1, 1, samples) at 24kHz

Source Dataset

• reach-vb/jenny_tts_dataset - Original Jenny TTS recordings

License

MIT (same as source dataset)