metadata
dataset_info:
features:
- name: speech
sequence: float32
- name: trascript
dtype: string
- name: answer
dtype: string
splits:
- name: train
num_bytes: 38175974440
num_examples: 123433
download_size: 37886689488
dataset_size: 38175974440
configs:
- config_name: default
data_files:
- split: train
path: data/train-*