metadata
dataset_info:
features:
- name: chunked_audio_filepath
dtype: audio
- name: text
dtype: string
- name: pred_text
dtype: string
- name: audio_filepath
dtype: string
- name: duration
dtype: float64
- name: alignment_score
dtype: float64
splits:
- name: hindi
num_bytes: 1385482988
num_examples: 39526
download_size: 1436911686
dataset_size: 1385482988
configs:
- config_name: default
data_files:
- split: hindi
path: data/hindi-*