metadata
dataset_info:
features:
- name: speaker_id
dtype: string
- name: audio
struct:
- name: bytes
dtype: binary
- name: path
dtype: string
- name: transcript
dtype: string
splits:
- name: train
num_bytes: 1444278974
num_examples: 9120
download_size: 1380935928
dataset_size: 1444278974
configs:
- config_name: default
data_files:
- split: train
path: data/train-*