metadata
dataset_info:
features:
- name: client_id
dtype: string
- name: audio
struct:
- name: array
sequence: float64
- name: sampling_rate
dtype: int64
- name: sentence
dtype: string
- name: condition_on_prev
sequence: int64
- name: whisper_transcript
dtype: string
splits:
- name: train
num_bytes: 223838055556
num_examples: 64504
download_size: 168603268559
dataset_size: 223838055556
configs:
- config_name: default
data_files:
- split: train
path: data/train-*