metadata
dataset_info:
features:
- name: json
struct:
- name: duration
dtype: float64
- name: speaker_name
dtype: string
- name: text
dtype: string
- name: __key__
dtype: string
- name: __url__
dtype: string
- name: text
dtype: string
- name: audio_tokens
sequence: int64
splits:
- name: train
num_bytes: 771825115
num_examples: 87135
- name: validation
num_bytes: 85362024
num_examples: 9682
download_size: 114546458
dataset_size: 857187139
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*