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:    ValueError
Message:      Dataset 'norm_fields_sca_max' has length 2 but expected 1
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 2840, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2373, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2398, 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 536, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 419, 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/hdf5/hdf5.py", line 76, in _generate_tables
                  num_rows = _check_dataset_lengths(h5, self.info.features)
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/hdf5/hdf5.py", line 353, in _check_dataset_lengths
                  raise ValueError(f"Dataset '{path}' has length {dset.shape[0]} but expected {num_rows}")
              ValueError: Dataset 'norm_fields_sca_max' has length 2 but expected 1

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.

PDE surrogate: the benchmark split

Companion data for the pde-surrogate task in T0-RSI/ai4sci-tasks. It carries only the files this benchmark's split creates, not a mirror of the upstream datasets.

subtree what
train/ the nine corpus files that carry training and held-out trajectories, truncated at the end of the validation range, with norm_* recomputed over the kept trajectories
eval/ the held-out tail of each of those files, under its original trajectory numbering

Every member keeps the sims group attributes (PDE, Fields, Fields Scheme, Dt, ...) and was opened with the task's own reader (environment/evaluation/pde_io.py) when it was built. Release pde-surrogate-split-v2 replaces pde-surrogate-split-v1 (revision 5bcba793), whose members were written without those attributes and which carried upstream's hyp.hdf5, a content copy of diff.hdf5; the task's split no longer uses that file.

Everything else the task uses is byte-identical to a pinned, public upstream file and is fetched from there: thuerey-group/pde-transformer-ape2d-full (MIT), polymathic-ai/turbulent_radiative_layer_2D and polymathic-ai/active_matter (CC-BY-4.0), and pdearena/NavierStokes-2D (MIT). Their licences and attribution apply unchanged; no relicensing is claimed here, and this repository does not duplicate their bytes.

Use

python3 environment/data/materialize.py DEST --from-release --subtree train   # agent host
python3 environment/data/materialize.py DEST --from-release                   # verifier host

release-manifest.json and payload-files.jsonl carry a SHA-256 per member; the task pins this repository by revision and manifest hash in environment/data/download-source.json.

Evaluation visibility

The task states: "All upstream data are public, so this is a benchmark-visible evaluation, not a secret test set." train/ and eval/ are separated so that an agent environment mounts only the former; together they reconstruct the upstream files exactly.

Downloads last month
59