Dataset Viewer
Auto-converted to Parquet
The dataset viewer is not available for this split.
Cannot load the dataset split (in streaming mode) to extract the first rows.
Error code:   StreamingRowsError
Exception:    CastError
Message:      Couldn't cast
json: struct<entities: struct<boxes: list<item: null>, goal: struct<bbox: struct<center_x: int64, center_y (... 538 chars omitted)
  child 0, entities: struct<boxes: list<item: null>, goal: struct<bbox: struct<center_x: int64, center_y: int64, height:  (... 233 chars omitted)
      child 0, boxes: list<item: null>
          child 0, item: null
      child 1, goal: struct<bbox: struct<center_x: int64, center_y: int64, height: int64, width: int64, x: int64, y: int6 (... 42 chars omitted)
          child 0, bbox: struct<center_x: int64, center_y: int64, height: int64, width: int64, x: int64, y: int64>
              child 0, center_x: int64
              child 1, center_y: int64
              child 2, height: int64
              child 3, width: int64
              child 4, x: int64
              child 5, y: int64
          child 1, pixel_pos: struct<x: int64, y: int64>
              child 0, x: int64
              child 1, y: int64
      child 2, player: struct<bbox: struct<center_x: int64, center_y: int64, height: int64, width: int64, x: int64, y: int6 (... 42 chars omitted)
          child 0, bbox: struct<center_x: int64, center_y: int64, height: int64, width: int64, x: int64, y: int64>
              child 0, center_x: int64
              child 1, center_y: int64
              child 2, height: int64
              child 3, width: int64
              child 4, x: int64
              child 5, y: int64
          child 1, pixel_pos: struct<x: int64, y: int64>
              child 0, x: int64
              child 1, y: int64
  child 1, game_type: string
  child 2, metadata: struct<road_width: int64, segments: list<item: list<item: list<item: double>>>, solution_path: list< (... 64 chars omitted)
      child 0, road_width: int64
      child 1, segments: list<item: list<item: list<item: double>>>
          child 0, item: list<item: list<item: double>>
              child 0, item: list<item: double>
                  child 0, item: double
      child 2, solution_path: list<item: list<item: double>>
          child 0, item: list<item: double>
              child 0, item: double
      child 3, solution_segments: list<item: int64>
          child 0, item: int64
  child 3, render: struct<cell_size: int64, image_height: int64, image_width: int64>
      child 0, cell_size: int64
      child 1, image_height: int64
      child 2, image_width: int64
  child 4, version: string
__key__: string
__url__: string
png: null
to
{'png': Image(mode=None, decode=True), '__key__': Value('string'), '__url__': Value('string')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 99, in get_rows_or_raise
                  return get_rows(
                         ^^^^^^^^^
                File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
                  return func(*args, **kwargs)
                         ^^^^^^^^^^^^^^^^^^^^^
                File "/src/services/worker/src/worker/utils.py", line 77, in get_rows
                  rows_plus_one = list(itertools.islice(ds, rows_max_number + 1))
                                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2431, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 1952, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 1984, in _iter_arrow
                  pa_table = cast_table_to_features(pa_table, self.features)
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2192, in cast_table_to_features
                  raise CastError(
              datasets.table.CastError: Couldn't cast
              json: struct<entities: struct<boxes: list<item: null>, goal: struct<bbox: struct<center_x: int64, center_y (... 538 chars omitted)
                child 0, entities: struct<boxes: list<item: null>, goal: struct<bbox: struct<center_x: int64, center_y: int64, height:  (... 233 chars omitted)
                    child 0, boxes: list<item: null>
                        child 0, item: null
                    child 1, goal: struct<bbox: struct<center_x: int64, center_y: int64, height: int64, width: int64, x: int64, y: int6 (... 42 chars omitted)
                        child 0, bbox: struct<center_x: int64, center_y: int64, height: int64, width: int64, x: int64, y: int64>
                            child 0, center_x: int64
                            child 1, center_y: int64
                            child 2, height: int64
                            child 3, width: int64
                            child 4, x: int64
                            child 5, y: int64
                        child 1, pixel_pos: struct<x: int64, y: int64>
                            child 0, x: int64
                            child 1, y: int64
                    child 2, player: struct<bbox: struct<center_x: int64, center_y: int64, height: int64, width: int64, x: int64, y: int6 (... 42 chars omitted)
                        child 0, bbox: struct<center_x: int64, center_y: int64, height: int64, width: int64, x: int64, y: int64>
                            child 0, center_x: int64
                            child 1, center_y: int64
                            child 2, height: int64
                            child 3, width: int64
                            child 4, x: int64
                            child 5, y: int64
                        child 1, pixel_pos: struct<x: int64, y: int64>
                            child 0, x: int64
                            child 1, y: int64
                child 1, game_type: string
                child 2, metadata: struct<road_width: int64, segments: list<item: list<item: list<item: double>>>, solution_path: list< (... 64 chars omitted)
                    child 0, road_width: int64
                    child 1, segments: list<item: list<item: list<item: double>>>
                        child 0, item: list<item: list<item: double>>
                            child 0, item: list<item: double>
                                child 0, item: double
                    child 2, solution_path: list<item: list<item: double>>
                        child 0, item: list<item: double>
                            child 0, item: double
                    child 3, solution_segments: list<item: int64>
                        child 0, item: int64
                child 3, render: struct<cell_size: int64, image_height: int64, image_width: int64>
                    child 0, cell_size: int64
                    child 1, image_height: int64
                    child 2, image_width: int64
                child 4, version: string
              __key__: string
              __url__: string
              png: null
              to
              {'png': Image(mode=None, decode=True), '__key__': Value('string'), '__url__': Value('string')}
              because column names don't match

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

VR-Bench Dataset

VR-Bench is a benchmark dataset for evaluating spatial reasoning capabilities of Vision-Language Models (VLMs) and Video Generation Models.

Dataset Structure

The dataset is split into two subsets:

dataset_VR_split/
├── train/          # Training set (96 cases)
│   ├── maze/
│   ├── maze3d/
│   ├── pathfinder/
│   ├── sokoban/
│   └── trapfield/
└── eval/           # Evaluation set (24 cases)
    ├── maze/
    ├── maze3d/
    ├── pathfinder/
    ├── sokoban/
    └── trapfield/

Each game directory contains:

  • images/: Initial state images (PNG)
  • states/: Game state metadata (JSON)
  • videos/: Solution trajectory videos (MP4)

Games

  • Maze: 2D grid-based navigation with walls
  • TrapField: 2D grid-based navigation with traps
  • Sokoban: Box-pushing puzzle game
  • PathFinder: Irregular maze with curved paths
  • Maze3D: 3D maze with vertical navigation

Usage

For VLM Evaluation

from datasets import load_dataset

dataset = load_dataset("your-username/VR-Bench")
train_data = dataset["train"]
eval_data = dataset["eval"]

For Video Model Evaluation

Each video file shows the optimal solution trajectory for the corresponding game state.

Citation

If you use this dataset, please cite:

@article{yang2025vrbench,
      title={Reasoning via Video: The First Evaluation of Video Models' Reasoning Abilities through Maze-Solving Tasks}, 
      author={Cheng Yang and Haiyuan Wan and Yiran Peng and Xin Cheng and Zhaoyang Yu and Jiayi Zhang and Junchi Yu and Xinlei Yu and Xiawu Zheng and Dongzhan Zhou and Chenglin Wu},
      journal={arXiv preprint arXiv:2511.15065},
      year={2025}
}

License

MIT License

Downloads last month
75