Dataset Viewer
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: Column 1 named env_params expected length 8192 but got length 32
Traceback: 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/hdf5/hdf5.py", line 87, in _generate_tables
pa_table = _recursive_load_arrays(h5, self.info.features, start, end)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/hdf5/hdf5.py", line 283, in _recursive_load_arrays
return pa.Table.from_pydict(batch_dict)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "pyarrow/table.pxi", line 1985, in pyarrow.lib._Tabular.from_pydict
File "pyarrow/table.pxi", line 6401, in pyarrow.lib._from_pydict
File "pyarrow/table.pxi", line 4912, in pyarrow.lib.Table.from_arrays
File "pyarrow/table.pxi", line 4256, in pyarrow.lib.Table.validate
File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
pyarrow.lib.ArrowInvalid: Column 1 named env_params expected length 8192 but got length 32Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
disco-combined
Trajectories used in the ICML 2026 paper Test-time Generalization for Physics through Neural Operator Splitting (Serrano, Han, Oyallon, Ho, Morel).
Three single-physics families (1D, 256 spatial points, 250 temporal snapshots over T=4s on a periodic domain of length L=16):
- EULER — pure nonlinear advection (Burgers-like), parameter α
- HEAT — pure diffusion, parameter β
- DISP — pure dispersion (linear), parameter γ
Each family has train, valid, and test splits.
Files
| File | Equation | Split |
|---|---|---|
combined_EULER_train.h5 |
nonlinear advection | train |
combined_EULER_valid.h5 |
nonlinear advection | valid |
combined_EULER_test.h5 |
nonlinear advection | test |
combined_HEAT_train.h5 |
diffusion | train |
combined_HEAT_valid.h5 |
diffusion | valid |
combined_HEAT_test.h5 |
diffusion | test |
combined_DISP_train.h5 |
dispersion | train |
combined_DISP_valid.h5 |
dispersion | valid |
combined_DISP_test.h5 |
dispersion | test |
Format
Each HDF5 file contains:
trajectories: shape(N, T, C, *spatial), float32env_id: shape(N,), int64 — environment index for each trajectoryenv_params/*: optional metadata mappingenv_idback to PDE coefficients
Filename mapping for the reference code
The training and test scripts in
LouisSerrano/neural-operator-splitting
expect a different naming convention (e.g. E_EULER_train_envsize64.h5). The
underlying data is identical — only the filenames differ. Use the snippet below
to download the files and symlink them to the names the configs reference:
| HF filename | Code-config filename |
|---|---|
combined_EULER_train.h5 |
E_EULER_train_envsize64.h5 |
combined_EULER_valid.h5 |
E_EULER_valid.h5 |
combined_EULER_test.h5 |
E_EULER_test.h5 |
combined_HEAT_train.h5 |
E_HEAT_train_envsize64.h5 |
combined_HEAT_valid.h5 |
E_HEAT_valid.h5 |
combined_HEAT_test.h5 |
E_HEAT_test.h5 |
combined_DISP_train.h5 |
E_DISP_train_envsize64.h5 |
combined_DISP_valid.h5 |
E_DISP_valid.h5 |
combined_DISP_test.h5 |
E_DISP_test.h5 |
import os
from huggingface_hub import hf_hub_download
REPO = "sogeeking/disco-combined"
TARGET_DIR = "./datasets/combined_equation"
os.makedirs(TARGET_DIR, exist_ok=True)
MAPPING = {
"combined_EULER_train.h5": "E_EULER_train_envsize64.h5",
"combined_EULER_valid.h5": "E_EULER_valid.h5",
"combined_EULER_test.h5": "E_EULER_test.h5",
"combined_HEAT_train.h5": "E_HEAT_train_envsize64.h5",
"combined_HEAT_valid.h5": "E_HEAT_valid.h5",
"combined_HEAT_test.h5": "E_HEAT_test.h5",
"combined_DISP_train.h5": "E_DISP_train_envsize64.h5",
"combined_DISP_valid.h5": "E_DISP_valid.h5",
"combined_DISP_test.h5": "E_DISP_test.h5",
}
for hf_name, local_name in MAPPING.items():
src = hf_hub_download(repo_id=REPO, filename=hf_name, repo_type="dataset")
dst = os.path.join(TARGET_DIR, local_name)
if not os.path.exists(dst):
os.symlink(src, dst)
Loading directly
from huggingface_hub import hf_hub_download
local = hf_hub_download(
repo_id="sogeeking/disco-combined",
filename="combined_EULER_train.h5",
repo_type="dataset",
)
Code: https://github.com/LouisSerrano/neural-operator-splitting
- Downloads last month
- 105