metadata
dataset_info:
config_name: train
features:
- name: id
dtype: string
- name: text
dtype: string
- name: input_features
dtype: image
- name: condition_on_prev
sequence: int64
- name: whisper_transcript
sequence: int64
splits:
- name: train
num_bytes: 4009681656693
num_examples: 2607550
download_size: 2515115685452
dataset_size: 4009681656693
configs:
- config_name: train
data_files:
- split: train
path: train/train-*