Dataset Viewer
Duplicate
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:    ArrowInvalid
Message:      JSON parse error: The document is empty.
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 280, in _generate_tables
                  df = pandas_read_json(f)
                       ^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 34, in pandas_read_json
                  return pd.read_json(path_or_buf, **kwargs)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/io/json/_json.py", line 815, in read_json
                  return json_reader.read()
                         ^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/io/json/_json.py", line 1014, in read
                  obj = self._get_object_parser(self.data)
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/io/json/_json.py", line 1040, in _get_object_parser
                  obj = FrameParser(json, **kwargs).parse()
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/io/json/_json.py", line 1176, in parse
                  self._parse()
                File "/usr/local/lib/python3.12/site-packages/pandas/io/json/_json.py", line 1391, in _parse
                  self.obj = DataFrame(
                             ^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/core/frame.py", line 778, in __init__
                  mgr = dict_to_mgr(data, index, columns, dtype=dtype, copy=copy, typ=manager)
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/core/internals/construction.py", line 503, in dict_to_mgr
                  return arrays_to_mgr(arrays, columns, index, dtype=dtype, typ=typ, consolidate=copy)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/core/internals/construction.py", line 114, in arrays_to_mgr
                  index = _extract_index(arrays)
                          ^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/core/internals/construction.py", line 677, in _extract_index
                  raise ValueError("All arrays must be of the same length")
              ValueError: All arrays must be of the same length
              
              During handling of the above exception, another exception occurred:
              
              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 2690, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2227, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2251, in _iter_arrow
                  for key, pa_table in self.ex_iterable._iter_arrow():
                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 494, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 384, in _iter_arrow
                  for key, pa_table in self.generate_tables_fn(**gen_kwags):
                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 283, in _generate_tables
                  raise e
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 246, in _generate_tables
                  pa_table = paj.read_json(
                             ^^^^^^^^^^^^^^
                File "pyarrow/_json.pyx", line 342, in pyarrow._json.read_json
                File "pyarrow/error.pxi", line 155, in pyarrow.lib.pyarrow_internal_check_status
                File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
              pyarrow.lib.ArrowInvalid: JSON parse error: The document is empty.

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.

MisCue3D

MisCue3D is a diagnostic dataset for studying context dependence in visually grounded 3D pose estimation. The dataset is designed to test whether a model's predicted pose, depth, or scale changes when the target object itself is held fixed but the surrounding visual context is changed.

This repository contains rendered observations and metadata from controlled Unreal Engine experiments. It does not contain Unreal Engine project files, .uasset files, source meshes, source textures, or marketplace/source 3D assets.

The dataset is intended for diagnostic evaluation.

Context families

MisCue3D contains three controlled context families.

1. Scale referent

A familiar referent object is rendered near the target object. The referent object's scale is varied while the target object and camera are held fixed.

Conditions include:

  • 1x: normal referent scale
  • 0.5x: smaller referent scale
  • 2x: larger referent scale

2. Support plausibility

The target object is rendered either with a plausible support surface or without support.

Conditions include:

  • with_support: target has a plausible support context
  • without_support: target appears unsupported or floating

3. Insertion vs. no-container

An elongated target object is rendered either inserted into a container that occludes part of it, or rendered alone with the same target pose.

Conditions include:

  • with_container: target is inserted into a container
  • without_container: container is removed while the target pose is preserved

Repository structure

This repository contains:

  • Insertion_no_container/
    Rendered examples and metadata for the insertion versus no-container context family.

  • NYC_East_Village_floating_object/
    Rendered examples and metadata for the support plausibility context family.

  • NYC_East_Village_scale_referent/
    Rendered examples and metadata for the scale-referent context family.

  • summary.json
    Dataset-level summary file.

  • insertion_no_container_sampled_by_us_object_paths_params.csv
    Sampled object/path parameter file for the insertion/no-container family.

  • data_manifest.jsonl
    File-level manifest listing rendered image files and their relative paths.

Rendered file types

The dataset may contain the following rendered image types:

  • lit.png
    RGB rendered image.

  • seg.png
    Segmentation image.

  • normal.png
    Surface-normal rendering.

The exact directory layout follows the context family, camera, and condition structure used during data generation.

What is included

This repository includes:

  • rendered RGB images
  • segmentation renderings
  • normal renderings
  • metadata files
  • file manifests
  • condition labels and experiment organization files

What is not included

This repository does not redistribute:

  • Unreal Engine project files
  • Unreal Engine map/source scene files
  • .uasset files
  • raw 3D meshes
  • raw textures or materials
  • marketplace/source 3D assets

The released files are rendered outputs and metadata only.

Intended use

MisCue3D is intended for research on:

  • 3D pose estimation
  • visually grounded 3D reconstruction
  • context dependence in single-image 3D models
  • robustness and diagnostic evaluation
  • controlled evaluation of pose, depth, and scale predictions

It is not intended as a general-purpose object recognition dataset or a source asset repository.

Limitations

MisCue3D is synthetic and generated from controlled rendering experiments. Its context families are designed to isolate specific visual cues, so the distribution is intentionally diagnostic rather than naturally sampled. Results on MisCue3D should therefore be interpreted as evidence of controlled context sensitivity, not as a direct estimate of real-world deployment performance.

The dataset reflects the selected object categories, rendered environments, camera configurations, and context interventions used during generation. Models may behave differently under other rendering engines, object distributions, lighting conditions, or real-image data.

License

Rendered images, annotations, metadata, and manifest files in this repository are released under the Creative Commons Attribution 4.0 International License (CC BY 4.0).

Third-party source assets, Unreal Engine project files, raw meshes, raw textures, materials, and marketplace/source assets are not redistributed in this repository and remain subject to their original licenses.

Citation

Citation information will be added after the associated anonymous submission is de-anonymized.

Downloads last month
4