metadata

configs:
  - config_name: random
    default: true
    data_files:
      - split: train
        path: random/train.jsonl
      - split: test
        path: random/test.jsonl
      - split: dev
        path: random/dev.jsonl
  - config_name: zeroshot
    data_files:
      - split: train
        path: zeroshot/train.jsonl
      - split: test
        path: zeroshot/test.jsonl
      - split: dev
        path: zeroshot/dev.jsonl
license: apache-2.0
language:
  - en
tags:
  - visual reasoning
  - grounded chat
  - visual grounding
size_categories:
  - 1K<n<10K

Grounded Visual Spatial Reasoning

Code for generating the annotations can be found here: github.com

Dataset Summary

This dataset extends the Visual Spatial Reasoning (VSR) dataset with visual grounding annotations: each caption is annotated with COCO-category object mentions, their positions , and corresponding bounding boxes in the image.

Data instance

Each sample instance has the following structure:

Field	Type	Description
`image_file`	`string`	COCO-style image filename
`image_link`	`string`	Direct COCO image URL
`width`	`int`	Image width
`height`	`int`	Image height
`caption`	`string`	Caption with two COCO-category object mentions
`label`	`bool`	Label from VSR original dataset
`relation`	`string`	Spatial relation
`ref_exp.labels`	`list[string]`	List of object labels from COCO categories
`ref_exp.label_positions`	`list[list[int]]`	Position (start, end) of each label in caption sentence
`ref_exp.bboxes`	`list[list[float]]`	Bounding boxes (`[x, y, w, h]` format)

Download Images

To download the images, follow the instructions from the VSR official GitHub repo.

Citation

If you use this dataset, please cite the original Visual Spatial Reasoning paper:

@article{Liu2022VisualSR,
  title={Visual Spatial Reasoning},
  author={Fangyu Liu and Guy Edward Toh Emerson and Nigel Collier},
  journal={Transactions of the Association for Computational Linguistics},
  year={2023},
}