metadata
dataset_info:
config_name: sw
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: 8071073313.528
num_examples: 9449
- name: validation
num_bytes: 2209317476.39
num_examples: 2597
- name: test
num_bytes: 2202385836.342
num_examples: 2602
download_size: 11612307194
dataset_size: 12482776626.259998
configs:
- config_name: sw
data_files:
- split: train
path: sw/train-*
- split: validation
path: sw/validation-*
- split: test
path: sw/test-*