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:    CastError
Message:      Couldn't cast
id: string
video_sha256: string
prompt_policy: string
source_video: string
prompt_sha256: string
reference_image_sha256: string
status: string
fps: double
bytes: int64
frames: int64
width: int64
height: int64
to
{'frames': Value('int64'), 'fps': Value('float64'), 'width': Value('int64'), 'height': Value('int64'), 'bytes': Value('int64')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 149, in get_rows_or_raise
                  return get_rows(
                      dataset=dataset,
                  ...<4 lines>...
                      column_names=column_names,
                  )
                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 129, in get_rows
                  rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
                File "/src/services/worker/src/worker/utils.py", line 544, in safe_iter
                  yield from ds.decode(False) if ds.features else ds
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 3052, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2559, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2584, in _iter_arrow
                  for key, pa_table in self.ex_iterable._iter_arrow():
                                       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 558, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 440, in _iter_arrow
                  for key, pa_table in self.generate_tables_fn(**gen_kwags):
                                       ~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 361, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2381, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2309, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              id: string
              video_sha256: string
              prompt_policy: string
              source_video: string
              prompt_sha256: string
              reference_image_sha256: string
              status: string
              fps: double
              bytes: int64
              frames: int64
              width: int64
              height: int64
              to
              {'frames': Value('int64'), 'fps': Value('float64'), 'width': Value('int64'), 'height': Value('int64'), 'bytes': Value('int64')}
              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.

PerfectPhysics Sliding: generated videos and tracking

126 generated videos: seven models × three sliding reference scenes × plain/enhanced prompts × seeds 45, 46, 47.

Models: Wan 2.2, Looking Ahead, DreamX-Creator, MAGI-1 4.5B, Cosmos 3 Nano, MiniMax H3 Base and LTX-2.5. GEN3C is excluded from this experiment.

Scenes: block_wood_25_0, block_plastic_25_0, and block_sp3k_25_0 (fine sandpaper), selected from worldbenchmark/PerfectPhysics, revision 47899855c8c285bd672539f74a6c6ed879bdfb8f.

Folder layout

<model>/<plain|enhanced>/<scene>/seed-<45|46|47>/
  video.mp4               complete generated video
  clip.json               identities, hashes, relative paths and processing status
  generation/             conditioning images, exact prompt, settings and provenance
  tracking/               SAM2 masks, overlay, raw tracks.csv and tracker metadata
  trajectories/raw.csv    convenient copy of the complete raw tracker table
  review-evidence/        sampled overlays, context sheet and frame inventory

Shared folders: qc/ contains mechanical tracking flags; references/ contains conditioning inputs, initialization annotations and reference inspection artifacts; provenance/ records prompt adaptations, source identities, settings and workflow scripts. manifest.json indexes all 126 clips; checksums.json records file sizes and SHA-256 hashes.

Conditioning and prompts

Generation uses the first reference frame only. The original 1280×720 scene is letterboxed into a 1280×1024 canvas (152 pixels above and below), preserving the full scene. Model-specific resizing is recorded with each output. Plain and enhanced wording comes from the Morpheus sliding_book prompts, with the object changed to a steel block and the surface description matched to wood, plastic or fine sandpaper. Exact source/adapted texts are in provenance/prompt-adaptations.json. Frame rate, duration and resolution can differ by model; consult each clip's recorded metadata.

Tracking status and use

SAM2 tracks the moving block from first-frame positive/negative initialization clicks. tracking/masks.npz stores masks, tracking/overlay.mp4 shows the propagated segmentation, and the CSV records frame, object ID, image-plane centroid and mask area. Missing detections retain the timeline with missing coordinates. These are image-plane observations, not metric 3D trajectories.

This snapshot contains raw, unreviewed tracks and mechanical QC evidence. Overlays are ready for visual review; missing masks, wrong-object tracking and other errors may remain. No reviewed/screened trajectories, final statistical scores or comparison figures are claimed in this release. Nonempty masks and successfully decoded videos do not establish physical correctness. Reference clip timing must be verified before time-aligned scoring comparisons.

Original source videos remain available from PerfectPhysics. This release does not assign a new blanket license to source material or model-generated outputs; consult the source dataset and individual model terms.

Downloads last month
19