Datasets:

SynDataLab-JA-Refs
/

irodori-refs-10k

license: apache-2.0
task_categories:
  - text-to-speech
language:
  - ja
size_categories:
  - 10K<n<100K
dataset_info:
  features:
    - name: audio
      dtype:
        audio:
          sampling_rate: 48000
    - name: text
      dtype: string
    - name: speaker_id
      dtype: string
  splits:
    - name: train
      num_examples: 10000
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*.parquet

Irodori TTS Reference Voices (10K)

10,000 synthetic Japanese reference voices generated with the Irodori-TTS-500M-v2-VoiceDesign model from voice-design captions (no reference audio — no_ref=True).

Each row is one unique speaker.

Columns

column	type	description
`audio`	Audio(48kHz mono)	reference waveform
`text`	string	Japanese utterance with emoji prosody cues
`speaker_id`	string	`speaker_00001` … `speaker_10000`

Related datasets

Clones (2.99M) — SynDataLab/irodori-clones-3m: 299 additional utterances generated in each reference voice using the main Irodori-TTS-500M-v2 model with latents encoded from these reference audios. Join on speaker_id to link each clone to its reference.

Generation

Model: Aratako/Irodori-TTS-500M-v2-VoiceDesign
Codec: Aratako/Semantic-DACVAE-Japanese-32dim
Precision: bf16, num_steps=40, truncation_factor=0.9, rescale_k=0.96, rescale_sigma=3.0, cfg_scale_text=3.0, cfg_scale_caption=3.0