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: 762431271
num_examples: 222
download_size: 573846537
dataset_size: 762431271
configs:
- config_name: default
data_files:
- split: train
path: data/train-*