metadata
dataset_info:
features:
- name: key_id
dtype: string
- name: audio
sequence: float64
- name: text
dtype: string
- name: audio_language
dtype: string
- name: speaker_id
dtype: string
- name: sample_rate
dtype: int64
splits:
- name: train
num_bytes: 3107675249
num_examples: 4994
download_size: 749265937
dataset_size: 3107675249
configs:
- config_name: default
data_files:
- split: train
path: data/train-*