metadata
dataset_info:
features:
- name: n_frames
dtype: uint32
- name: generated_text
dtype: string
- name: generated_tokens
list: int64
- name: all_frames
list: binary
- name: all_captions
list:
- name: c
dtype: string
- name: i
dtype: int64
- name: oi
dtype: int64
- name: p
dtype: int64
splits:
- name: train
num_bytes: 1179483519
num_examples: 1024
download_size: 1168233307
dataset_size: 1179483519
configs:
- config_name: default
data_files:
- split: train
path: data/train-*