---
dataset_info:
  features:
  - name: index
    dtype: string
  - name: datasetname
    dtype: string
  - name: audio
    dtype: audio
  - name: duration
    dtype: float64
  - name: audio_len
    dtype: float64
  - name: text
    dtype: string
  - name: raw_text
    list: string
  - name: split
    dtype: string
  - name: file_name
    dtype: string
  - name: text_ja
    dtype: string
  splits:
  - name: train
    num_bytes: 38981921523
    num_examples: 19690
  download_size: 37199176449
  dataset_size: 38981921523
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
task_categories:
- audio-classification
- text-generation
language:
- ja
---

[Qwen/Qwen3-4B-Instruct-2507](https://huggingface.co/Qwen/Qwen3-4B-Instruct-2507)を使用して[Clotho](https://zenodo.org/records/3490684)を日本語に翻訳したデータです。
ライセンスは元データに従います。