metadata
dataset_info:
config_name: hi
features:
- name: path
dtype: string
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: sentence
dtype: string
- name: condition_on_prev
sequence: int64
- name: whisper_transcript
dtype: string
splits:
- name: train
num_bytes: 622094627
num_examples: 704
- name: validation
num_bytes: 349393253
num_examples: 396
- name: test
num_bytes: 488626356
num_examples: 561
download_size: 1366638308
dataset_size: 1460114236
configs:
- config_name: hi
data_files:
- split: train
path: hi/train-*
- split: validation
path: hi/validation-*
- split: test
path: hi/test-*