metadata
dataset_info:
features:
- name: image
dtype: image
- name: sequence_name
dtype: string
- name: frame_id
dtype: int32
- name: objects
struct:
- name: bbox
list:
list: int32
length: 4
- name: category
list: int32
- name: target_id
list: int32
splits:
- name: train
num_bytes: 2757497636
num_examples: 24201
- name: validation
num_bytes: 899199835
num_examples: 2846
- name: test_dev
num_bytes: 3742528529
num_examples: 6635
- name: test_challenge
num_bytes: 3862990015
num_examples: 6333
download_size: 14989495720
dataset_size: 11262216015
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test_dev
path: data/test_dev-*
- split: test_challenge
path: data/test_challenge-*