---
dataset_info:
- config_name: captioning
  features:
  - name: image
    dtype: image
  - name: image_path
    dtype: string
  - name: prompts
    dtype: string
  - name: task
    dtype: string
  - name: ground_truth
    dtype: string
  - name: image_name
    dtype: string
  - name: question_id
    dtype: int32
  splits:
  - name: train
    num_bytes: 392144290
    num_examples: 434
  download_size: 391182976
  dataset_size: 392144290
- config_name: ref_det
  features:
  - name: image
    dtype: image
  - name: image_path
    dtype: string
  - name: prompts
    dtype: string
  - name: task
    dtype: string
  - name: ground_truth
    dtype: string
  - name: image_name
    dtype: string
  - name: question_id
    dtype: int32
  splits:
  - name: train
    num_bytes: 1129342745
    num_examples: 1313
  download_size: 1130621716
  dataset_size: 1129342745
- config_name: ref_seg
  features:
  - name: image
    dtype: image
  - name: mask
    dtype: image
  - name: image_path
    dtype: string
  - name: prompts
    dtype: string
  - name: task
    dtype: string
  - name: ground_truth
    dtype: string
  - name: image_name
    dtype: string
  - name: question_id
    dtype: int32
  splits:
  - name: train
    num_bytes: 1223957311
    num_examples: 530
  download_size: 1221308027
  dataset_size: 1223957311
- config_name: single
  features:
  - name: image
    dtype: image
  - name: image_path
    dtype: string
  - name: prompts
    dtype: string
  - name: options_list
    dtype: string
  - name: options
    dtype: string
  - name: task
    dtype: string
  - name: ground_truth
    dtype: string
  - name: ground_truth_option
    dtype: string
  - name: image_name
    dtype: string
  - name: question_id
    dtype: int32
  - name: cls_description
    dtype: string
  splits:
  - name: train
    num_bytes: 1436686319
    num_examples: 3211
  download_size: 1650341032
  dataset_size: 1436686319
- config_name: temporal
  features:
  - name: images
    list: binary
  - name: num_images
    dtype: int32
  - name: image_path
    dtype: string
  - name: prompts
    dtype: string
  - name: options_list
    dtype: string
  - name: image_name
    dtype: string
  - name: task
    dtype: string
  - name: ground_truth
    dtype: string
  - name: ground_truth_option
    dtype: string
  - name: options
    dtype: string
  - name: question_id
    dtype: int32
  splits:
  - name: train
    num_bytes: 1767293990
    num_examples: 1713
  download_size: 1745747329
  dataset_size: 1767293990
configs:
- config_name: captioning
  data_files:
  - split: train
    path: captioning/train-*
- config_name: ref_det
  data_files:
  - split: train
    path: ref_det/train-*
- config_name: ref_seg
  data_files:
  - split: train
    path: ref_seg/train-*
- config_name: single
  data_files:
  - split: train
    path: single/train-*
- config_name: temporal
  data_files:
  - split: train
    path: temporal/train-*
---