metadata
dataset_info:
features:
- name: path
dtype: string
- name: shape
sequence: int32
- name: volume
dtype: int32
- name: blocks
sequence:
sequence:
sequence:
sequence: int32
- name: image
dtype: image
- name: block_counts
sequence:
sequence: int32
- name: descriptions
sequence: string
- name: descriptions_scores
sequence: int64
- name: top_description
dtype: string
splits:
- name: train
num_bytes: 54208291
num_examples: 655
download_size: 14813433
dataset_size: 54208291
configs:
- config_name: default
data_files:
- split: train
path: data/train-*