metadata
dataset_info:
features:
- name: wav_filename
dtype: string
- name: duration
dtype: float64
- name: text
dtype: string
- name: transcript
dtype: string
- name: levenshtein
dtype: float64
- name: speaker
dtype: int64
- name: file_size
dtype: int64
- name: num_words
dtype: int64
- name: gender
dtype: string
- name: accent
dtype: string
- name: audio
dtype: audio
splits:
- name: validation
num_bytes: 629252352.444
num_examples: 1158
- name: test
num_bytes: 645189430.752
num_examples: 1316
- name: train
num_bytes: 26324500255.68
num_examples: 73848
download_size: 102232814813
dataset_size: 27598942038.876
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*