metadata
dataset_info:
features:
- name: image_id
dtype: int32
- name: image
dtype:
array3_d:
shape:
- 800
- 600
- 3
dtype: uint8
- name: relationships_text
dtype: string
- name: objects
sequence: string
splits:
- name: train
num_bytes: 8884156
num_examples: 3
download_size: 3471472
dataset_size: 8884156
configs:
- config_name: default
data_files:
- split: train
path: data/train-*