metadata
dataset_info:
features:
- name: audio_id
dtype: string
- name: caption
dtype: string
- name: audio
dtype: audio
- name: timeline
dtype: image
- name: phrases
list:
- name: original_clip_id
dtype: int64
- name: phrase
dtype: string
- name: segments
list:
- name: end
dtype: float64
- name: start
dtype: float64
splits:
- name: train
num_bytes: 57166176827
num_examples: 100000
download_size: 57197822960
dataset_size: 57166176827
configs:
- config_name: default
data_files:
- split: train
path: data/train-*