metadata
dataset_info:
features:
- name: sent_ids
sequence: int64
- name: file_name
dtype: string
- name: ann_id
dtype: int64
- name: ref_id
dtype: int64
- name: image_id
dtype: int64
- name: split
dtype: string
- name: sentences
list:
- name: raw
dtype: string
- name: sent
dtype: string
- name: sent_id
dtype: int64
- name: tokens
sequence: string
- name: category_id
dtype: int64
- name: raw_anns
dtype: string
- name: raw_image_info
dtype: string
- name: raw_sentences
dtype: string
- name: image_path
dtype: string
- name: bbox
sequence: float64
- name: captions
sequence: string
- name: global_image_id
dtype: string
splits:
- name: train
num_bytes: 80612144.55392
num_examples: 42404
- name: testB
num_bytes: 3440901.3688
num_examples: 1810
- name: test
num_bytes: 3754574.698
num_examples: 1975
- name: validation
num_bytes: 7244903.37928
num_examples: 3811
download_size: 38427962
dataset_size: 95052524