Hugging Face's logo

Datasets:
SourajaKundu123
/
PhySimCode

Tasks:
Video-Text-to-Text
Modalities:
Video
Languages:
English
Size:
1K<n<10K
Tags:
physics
simulation
video
code-generation
multimodal
benchmark
License:
Dataset card Files
xet
 Community
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
seed: double
module: double
addendum: double
dedendum: double
tooth_frac: double
r_pulley: double
omega_max: double
t_ramp: double
x_axle: double
y_rack: double
rack_length: double
rack_height: double
x_rack_0: double
duration: double
fps: int64
N_top: int64
N_mid: int64
R_top: double
R_mid: double
y_top: double
y_mid: double
physical_law: struct<name: string, statement: string, formula: string>
  child 0, name: string
  child 1, statement: string
  child 2, formula: string
code_validation: struct<ode_source: string, energy_check: string, physical_plausibility: bool>
  child 0, ode_source: string
  child 1, energy_check: string
  child 2, physical_plausibility: bool
physical_observation: string
simulation: string
parameters: struct<seed: struct<value: double, unit: string, description: string>, module: struct<value: double, (... 1305 chars omitted)
  child 0, seed: struct<value: double, unit: string, description: string>
      child 0, value: double
      child 1, unit: string
      child 2, description: string
  child 1, module: struct<value: double, unit: string, description: string>
      child 0, value: double
      child 1, unit: string
      child 2, description: string
  child 2, addendum: struct<value: double, unit: string, description: string>
      child 0, value: double
      child 1, unit: string
      child 2, description: string
  child 3, dedendum: struct<value: double, unit: string, description: string>
      child 0, value: double
      child 1, unit: string

...
   child 2, description: string
  child 18, R_mid: struct<value: double, unit: string, description: string>
      child 0, value: double
      child 1, unit: string
      child 2, description: string
  child 19, y_top: struct<value: double, unit: string, description: string>
      child 0, value: double
      child 1, unit: string
      child 2, description: string
  child 20, y_mid: struct<value: double, unit: string, description: string>
      child 0, value: double
      child 1, unit: string
      child 2, description: string
derivation_steps: list<item: string>
  child 0, item: string
expected_behavior: list<item: string>
  child 0, item: string
physical_constants_derived: struct<omega_top_max_rad_per_s: double, omega_mid_max_rad_per_s: double, v_rack_max_m_per_s: double, (... 23 chars omitted)
  child 0, omega_top_max_rad_per_s: double
  child 1, omega_mid_max_rad_per_s: double
  child 2, v_rack_max_m_per_s: double
  child 3, t_disengage_s: double
numerical_method: struct<solver: string, tolerance: string, justification: string>
  child 0, solver: string
  child 1, tolerance: string
  child 2, justification: string
ode_system: struct<state_variables: list<item: string>, equations_latex: list<item: string>, equations_python: l (... 18 chars omitted)
  child 0, state_variables: list<item: string>
      child 0, item: string
  child 1, equations_latex: list<item: string>
      child 0, item: string
  child 2, equations_python: list<item: string>
      child 0, item: string
to
{'simulation': Value('string'), 'physical_observation': Value('string'), 'physical_law': {'name': Value('string'), 'statement': Value('string'), 'formula': Value('string')}, 'ode_system': {'state_variables': List(Value('string')), 'equations_latex': List(Value('string')), 'equations_python': List(Value('string'))}, 'derivation_steps': List(Value('string')), 'physical_constants_derived': {'omega_top_max_rad_per_s': Value('float64'), 'omega_mid_max_rad_per_s': Value('float64'), 'v_rack_max_m_per_s': Value('float64'), 't_disengage_s': Value('float64')}, 'numerical_method': {'solver': Value('string'), 'tolerance': Value('string'), 'justification': Value('string')}, 'code_validation': {'ode_source': Value('string'), 'energy_check': Value('string'), 'physical_plausibility': Value('bool')}, 'expected_behavior': List(Value('string')), 'parameters': {'seed': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'module': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'addendum': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'dedendum': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'tooth_frac': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'r_pulley': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'omega_max': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 't_ramp': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'x_axle': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'y_rack': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'rack_length': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'rack_height': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'x_rack_0': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'duration': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'fps': {'value': Value('int64'), 'unit': Value('string'), 'description': Value('string')}, 'N_top': {'value': Value('int64'), 'unit': Value('string'), 'description': Value('string')}, 'N_mid': {'value': Value('int64'), 'unit': Value('string'), 'description': Value('string')}, 'R_top': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'R_mid': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'y_top': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'y_mid': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}}}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 147, 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 127, 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 483, 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 2951, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2461, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2486, 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 547, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 430, 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 343, 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 2378, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              seed: double
              module: double
              addendum: double
              dedendum: double
              tooth_frac: double
              r_pulley: double
              omega_max: double
              t_ramp: double
              x_axle: double
              y_rack: double
              rack_length: double
              rack_height: double
              x_rack_0: double
              duration: double
              fps: int64
              N_top: int64
              N_mid: int64
              R_top: double
              R_mid: double
              y_top: double
              y_mid: double
              physical_law: struct<name: string, statement: string, formula: string>
                child 0, name: string
                child 1, statement: string
                child 2, formula: string
              code_validation: struct<ode_source: string, energy_check: string, physical_plausibility: bool>
                child 0, ode_source: string
                child 1, energy_check: string
                child 2, physical_plausibility: bool
              physical_observation: string
              simulation: string
              parameters: struct<seed: struct<value: double, unit: string, description: string>, module: struct<value: double, (... 1305 chars omitted)
                child 0, seed: struct<value: double, unit: string, description: string>
                    child 0, value: double
                    child 1, unit: string
                    child 2, description: string
                child 1, module: struct<value: double, unit: string, description: string>
                    child 0, value: double
                    child 1, unit: string
                    child 2, description: string
                child 2, addendum: struct<value: double, unit: string, description: string>
                    child 0, value: double
                    child 1, unit: string
                    child 2, description: string
                child 3, dedendum: struct<value: double, unit: string, description: string>
                    child 0, value: double
                    child 1, unit: string
              
              ...
                 child 2, description: string
                child 18, R_mid: struct<value: double, unit: string, description: string>
                    child 0, value: double
                    child 1, unit: string
                    child 2, description: string
                child 19, y_top: struct<value: double, unit: string, description: string>
                    child 0, value: double
                    child 1, unit: string
                    child 2, description: string
                child 20, y_mid: struct<value: double, unit: string, description: string>
                    child 0, value: double
                    child 1, unit: string
                    child 2, description: string
              derivation_steps: list<item: string>
                child 0, item: string
              expected_behavior: list<item: string>
                child 0, item: string
              physical_constants_derived: struct<omega_top_max_rad_per_s: double, omega_mid_max_rad_per_s: double, v_rack_max_m_per_s: double, (... 23 chars omitted)
                child 0, omega_top_max_rad_per_s: double
                child 1, omega_mid_max_rad_per_s: double
                child 2, v_rack_max_m_per_s: double
                child 3, t_disengage_s: double
              numerical_method: struct<solver: string, tolerance: string, justification: string>
                child 0, solver: string
                child 1, tolerance: string
                child 2, justification: string
              ode_system: struct<state_variables: list<item: string>, equations_latex: list<item: string>, equations_python: l (... 18 chars omitted)
                child 0, state_variables: list<item: string>
                    child 0, item: string
                child 1, equations_latex: list<item: string>
                    child 0, item: string
                child 2, equations_python: list<item: string>
                    child 0, item: string
              to
              {'simulation': Value('string'), 'physical_observation': Value('string'), 'physical_law': {'name': Value('string'), 'statement': Value('string'), 'formula': Value('string')}, 'ode_system': {'state_variables': List(Value('string')), 'equations_latex': List(Value('string')), 'equations_python': List(Value('string'))}, 'derivation_steps': List(Value('string')), 'physical_constants_derived': {'omega_top_max_rad_per_s': Value('float64'), 'omega_mid_max_rad_per_s': Value('float64'), 'v_rack_max_m_per_s': Value('float64'), 't_disengage_s': Value('float64')}, 'numerical_method': {'solver': Value('string'), 'tolerance': Value('string'), 'justification': Value('string')}, 'code_validation': {'ode_source': Value('string'), 'energy_check': Value('string'), 'physical_plausibility': Value('bool')}, 'expected_behavior': List(Value('string')), 'parameters': {'seed': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'module': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'addendum': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'dedendum': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'tooth_frac': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'r_pulley': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'omega_max': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 't_ramp': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'x_axle': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'y_rack': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'rack_length': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'rack_height': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'x_rack_0': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'duration': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'fps': {'value': Value('int64'), 'unit': Value('string'), 'description': Value('string')}, 'N_top': {'value': Value('int64'), 'unit': Value('string'), 'description': Value('string')}, 'N_mid': {'value': Value('int64'), 'unit': Value('string'), 'description': Value('string')}, 'R_top': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'R_mid': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'y_top': {'value': Value('float64'), 'unit': Value('string'), 'description': Value('string')}, 'y_mid': {'value': Value('float64'), 'unit': Value('string'), 'description': 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.

PhySimCode: A Benchmark and Evaluation Method for Physics Video to Code Generation

Website · Paper (arXiv, coming soon) · Code

Can a multimodal LLM watch a 2D or 3D physics simulation, work out the involved physics laws and parameter values, and write code from scratch that regenerates it?

This dataset is the PhySimCode evaluation set: 2,430 physics simulation videos, each paired with the Python code that rendered it, its true parameter values, and a step-by-step physics explanation (chain-of-thought). It is the set on which all models in the paper are evaluated.

Overview

  • 162 physics phenomena across 9 physical domains (rigid body, spring / wave, articulated, constrained, pendulum, real-life mixed, rolling, rope / granular, gravitation / fluid)
  • 15 samples per phenomenon, each with randomly sampled physical parameters
  • Two engines: analytic SciPy simulations (2D) and PyBullet rigid-body scenes (3D)
  • Drawn from the full PhySimCode corpus of 160,614 procedurally generated samples
Engine Folder Phenomena Samples Video
SciPy (2D) scipy/ 87 1,305 360×180, 10 FPS
PyBullet (3D) kubric/ 75 1,125 240×240, 10 FPS
Total 162 2,430

Structure 

scipy/<experiment>/<sample_id>/
    video.mp4            # input to the model
    simulation_code.py   # ground-truth code that renders video.mp4
    cot.json             # ground-truth chain-of-thought
    params.json          # sampled parameter values
kubric/<experiment>/<sample_id>/
    ...                  # same four files

Every cot.json contains simulation, physical_observation, physical_law (name, statement, LaTeX formula), ode_system, derivation_steps, numerical_method, code_validation, expected_behavior and parameters (value, unit and description for every parameter).

  • SciPy samples also include physical_constants_derived (quantities computed from the parameters).
  • PyBullet samples also include category, seed, modeling_assumptions, impulse_model, pybullet_numerical_model and visual_scene.

Task

Given only video.mp4 and the engine name (scipy or kubric), a model must output:

  1. a chain-of-thought JSON with simulation, physical_observation, physical_law and parameters, and
  2. a self-contained Python script that regenerates the video.

Models are scored on physical-law correctness and equivalence (three LLM judges), parameter recovery within ±20%, code compilation / execution / video output, CodeBLEU, and DINOv2 / VideoCLIP similarity between the regenerated and input videos. The prompt is given in Appendix A.4 of the paper, and the evaluation code is on GitHub.

Download 

pip install -U huggingface_hub
hf download SourajaKundu123/PhySimCode --repo-type dataset --local-dir benchmarking_data

from huggingface_hub import snapshot_download
snapshot_download("SourajaKundu123/PhySimCode", repo_type="dataset", local_dir="benchmarking_data")

Citation 

@misc{kundu2026physimcode,
  title  = {PhySimCode: A Benchmark and Evaluation Method for
            Physics Video to Code Generation},
  author = {Kundu, Souraja and Gupta, Aditya and Julin, Joel and
            Zhao, Yizhou and Xie, Liuyue and Jeni, Laszlo A.},
  year   = {2026},
  note   = {arXiv preprint (coming soon)}
}

Contact

Souraja Kundu, sourajak@cs.cmu.edu
Downloads last month
-
Total file size:
72.9 MB