metadata
dataset_info:
config_name: release3
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: text
dtype: string
- name: speaker_id
dtype: string
- name: gender
dtype:
class_label:
names:
'0': unknown
'1': female
'2': male
- name: file
dtype: string
- name: id
dtype: string
- name: whisper_transcript_unprompted
dtype: string
- name: whisper_transcript
dtype: string
splits:
- name: train
num_bytes: 52484152554.125
num_examples: 268263
- name: validation
num_bytes: 184679438
num_examples: 507
- name: test
num_bytes: 302513272.625
num_examples: 1155
download_size: 52650349441
dataset_size: 52971345264.75
configs:
- config_name: release3
data_files:
- split: train
path: release3/train-*
- split: validation
path: release3/validation-*
- split: test
path: release3/test-*