metadata
dataset_info:
features:
- name: image
dtype: image
- name: image_id
dtype: int32
- name: objects
struct:
- name: bbox
sequence:
sequence: float32
- name: categories
sequence: int64
splits:
- name: train
num_bytes: 8694969
num_examples: 100
- name: validation
num_bytes: 2547546
num_examples: 30
- name: test
num_bytes: 1474594
num_examples: 30
download_size: 12491262
dataset_size: 12717109
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*