metadata
dataset_info:
features:
- name: source_audio
dtype:
audio:
sampling_rate: 16000
- name: source_text
dtype: string
- name: source_lang
dtype: string
- name: target_text
dtype: string
- name: target_lang
dtype: string
- name: target_audio
dtype: audio
- name: global_index
dtype: int64
splits:
- name: train
num_bytes: 6866666640
num_examples: 61885
download_size: 6862908266
dataset_size: 6866666640
configs:
- config_name: default
data_files:
- split: train
path: data/train-*