Dataset Viewer
The dataset viewer is not available for this subset.
Cannot get the split names for the config 'default' of the dataset.
Exception:    SplitsNotFoundError
Message:      The split names could not be parsed from the dataset config.
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 291, in _generate_tables
                  pa_table = paj.read_json(
                      io.BytesIO(batch), read_options=paj.ReadOptions(block_size=block_size)
                  )
                File "pyarrow/_json.pyx", line 364, in pyarrow._json.read_json
                File "pyarrow/error.pxi", line 155, in pyarrow.lib.pyarrow_internal_check_status
                  return check_status(status)
                File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
                  raise convert_status(status)
              pyarrow.lib.ArrowInvalid: JSON parse error: Column() changed from object to string in row 0
              
              During handling of the above exception, another exception occurred:
              
              Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 286, in get_dataset_config_info
                  for split_generator in builder._split_generators(
                                         ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      StreamingDownloadManager(base_path=builder.base_path, download_config=download_config)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 101, in _split_generators
                  pa_table = next(iter(self._generate_tables(**splits[0].gen_kwargs, allow_full_read=False)))[1]
                             ~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 306, in _generate_tables
                  batch = json_encode_fields_in_json_lines(original_batch, json_field_paths)
                File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 111, in json_encode_fields_in_json_lines
                  examples = [ujson_loads(line) for line in original_batch.splitlines()]
                              ~~~~~~~~~~~^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 20, in ujson_loads
                  return pd.io.json.ujson_loads(*args, **kwargs)
                         ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
              ValueError: Expected object or value
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/split_names.py", line 68, in compute_split_names_from_streaming_response
                  for split in get_dataset_split_names(
                               ~~~~~~~~~~~~~~~~~~~~~~~^
                      path=dataset,
                      ^^^^^^^^^^^^^
                      config_name=config,
                      ^^^^^^^^^^^^^^^^^^^
                      token=hf_token,
                      ^^^^^^^^^^^^^^^
                  )
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 340, in get_dataset_split_names
                  info = get_dataset_config_info(
                      path,
                  ...<6 lines>...
                      **config_kwargs,
                  )
                File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 291, in get_dataset_config_info
                  raise SplitsNotFoundError("The split names could not be parsed from the dataset config.") from err
              datasets.inspect.SplitsNotFoundError: The split names could not be parsed from the dataset config.

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.

Delighting Benchmark Diverse-50

50 cases. Start with index.json; every case uses the same layout:

cases/<case_id>/
  gt_mesh.glb          # standalone GT mesh in benchmark input coordinates
  gt_conversion.json  # conversion and precision evidence
  source/             # original source assets, UVs and textures
  inputs/             # benchmark geometry, input images and preparation metadata
  gt_renders/         # original GT lit/unlit reference videos
  methods/
    paint3d/          # predictions and their input files
    hunyuan3d/         # predictions and their input files

Identical method inputs are consolidated under inputs/ (model.obj and input_image.png). Any method-specific inputs remain with that method. Model predictions (textured.glb, albedo_texture.png) are NOT GT. gt_unlit_render.mp4 is a rendered reference, not a UV atlas. Original reports live under legacy_metadata/; their paths refer to the previous layout. path_migration.json maps old paths to new paths. Older Hugging Face revisions retain the original directory layout.

GT provenance and precision

Canonical-30: 20 official ABO source models and 10 official Poly Haven 4K assets recovered by original asset ID. Original geometry/UV accessor bytes, material bindings, and image bytes are preserved. Only a root node transform aligns them to the packaged benchmark OBJ; all face corners agree within 1e-6. No resampling, remeshing, or UV regeneration. Corresponding official assets are verified; exact historical texture resolution/version is NOT proven. Poly Haven CC0, Powered by Poly Haven; ABO CC BY 4.0, Amazon.com / ABO team.

Diverse-50: source GSO OBJ UVs and original PNG bytes, using benchmark OBJ positions. Float32 conversion introduces up to ~2.98e-8 attribute rounding. Materials are reconstructed (metallic=0, roughness=1) and normals generated because source MTL/normals are absent. GSO CC BY 4.0 / Google LLC.

Texture sizes are recorded per case in index.json. Original resolution is preserved; no upsampling is used. A 4K render does not imply a 4K source texture. Camera normalization for reference videos remains documented in legacy reports.

Visual verification

See gallery.html, contact_sheet.jpg, and render_validation.json. Every case has a 1024×1024 direct-source / GLB base-color render comparison; selected cases also have 2048×2048 front and back images. These test conversion fidelity, not exact historical asset-version identity.

Khronos glTF validation: zero errors. Canonical preserves 66 warnings already present in official source assets (65 missing explicit tangent-space warnings and one image-feature warning); Diverse has zero warnings. Source and output validation reports are included.

Legacy video variants

The previous repository also contained top-level paint3d/ and hunyuan3d/ GT videos. Some differ in bytes from the canonical_outputs/ references. Identical copies are consolidated; differing legacy versions are retained under legacy_metadata/ with their original relative paths. The current reference videos are explicitly cases/<case_id>/gt_renders/, migrated from the original canonical_outputs/ package. No differing video is discarded or silently substituted.

Downloads last month
329