metadata
dataset_info:
features:
- name: image
dtype: image
- name: annotation
dtype: image
- name: semantic_class_to_id
struct:
- name: background
dtype: int64
- name: car
dtype: int64
- name: person
dtype: int64
splits:
- name: train
num_bytes: 74687009
num_examples: 1600
- name: validation
num_bytes: 15833241
num_examples: 330
download_size: 89615583
dataset_size: 90520250
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*