metadata
dataset_info:
features:
- name: __key__
dtype: string
- name: __url__
dtype: string
- name: audio
dtype: audio
- name: json
struct:
- name: caption
dtype: string
- name: channel_follower_count
dtype: int64
- name: duration_ms
dtype: int64
- name: id
dtype: string
- name: like_count
dtype: int64
- name: segment_filename
dtype: string
- name: segment_index
dtype: int64
- name: start_time_ms
dtype: int64
- name: title
dtype: string
- name: transcription
dtype: string
- name: type
dtype: string
- name: uploader
dtype: string
- name: uploader_id
dtype: string
- name: view_count
dtype: int64
- name: text
dtype: string
splits:
- name: train
num_bytes: 1790445957.066
num_examples: 14706
- name: test
num_bytes: 198938460.916
num_examples: 1634
download_size: 1673012277
dataset_size: 1989384417.9819999
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*