metadata
license: apache-2.0
dataset_info:
config_name: zh-TW
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: 691009749
num_examples: 766
- name: validation
num_bytes: 501044472
num_examples: 554
- name: test
num_bytes: 552470080
num_examples: 617
download_size: 1560691983
dataset_size: 1744524301
configs:
- config_name: zh-TW
data_files:
- split: train
path: zh-TW/train-*
- split: validation
path: zh-TW/validation-*
- split: test
path: zh-TW/test-*