An analytic theory of convolutional neural network inverse problems solvers
Paper • 2601.10334 • Published
Dataset Preview
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed because of a cast error
Error code: DatasetGenerationCastError
Exception: DatasetGenerationCastError
Message: An error occurred while generating the dataset
All the data files must have the same columns, but at some point there are 3 new columns ({'network_formula_psnr_db', 'network_formula_mse', 'normalized_negative_log_measurement_density'}) and 6 missing columns ({'task', 'network_first_call_seconds', 'psnr_network_vs_formula_db', 'psnr_network_vs_ground_truth_db', 'psnr_formula_vs_ground_truth_db', 'formula_seconds'}).
This happened while the csv dataset builder was generating data using
hf://datasets/StochasticRhapsody/analytic-mmse-repro-results/measurement_density_probe.csv (at revision 5a9a868afcb6b15c39fc16d61606e9b5bb4c72c1), ['hf://datasets/StochasticRhapsody/analytic-mmse-repro-results@5a9a868afcb6b15c39fc16d61606e9b5bb4c72c1/checkpoint_formula_grid.csv', 'hf://datasets/StochasticRhapsody/analytic-mmse-repro-results@5a9a868afcb6b15c39fc16d61606e9b5bb4c72c1/measurement_density_probe.csv']
Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1837, in _prepare_split_single
writer.write_table(table)
~~~~~~~~~~~~~~~~~~^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/arrow_writer.py", line 765, in write_table
self._write_table(pa_table, writer_batch_size=writer_batch_size)
~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/arrow_writer.py", line 773, in _write_table
pa_table = table_cast(pa_table, self._schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2369, in table_cast
return cast_table_to_schema(table, schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2297, in cast_table_to_schema
raise CastError(
...<3 lines>...
)
datasets.table.CastError: Couldn't cast
test_index: int64
sigma: double
normalized_negative_log_measurement_density: double
network_formula_mse: double
network_formula_psnr_db: double
mean_log_patch_mass_unnormalized: double
-- schema metadata --
pandas: '{"index_columns": [{"kind": "range", "name": null, "start": 0, "' + 1143
to
{'task': Value('string'), 'sigma': Value('float64'), 'test_index': Value('int64'), 'psnr_network_vs_formula_db': Value('float64'), 'psnr_formula_vs_ground_truth_db': Value('float64'), 'psnr_network_vs_ground_truth_db': Value('float64'), 'formula_seconds': Value('float64'), 'network_first_call_seconds': Value('float64'), 'mean_log_patch_mass_unnormalized': Value('float64')}
because column names don't match
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
~~~~~~~~~~~~~~~~~~~~~~~~~^
builder, max_dataset_size_bytes=max_dataset_size_bytes
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1683, in _prepare_split
for job_id, done, content in self._prepare_split_single(
~~~~~~~~~~~~~~~~~~~~~~~~~~^
gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
):
^
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1839, in _prepare_split_single
raise DatasetGenerationCastError.from_cast_error(
...<4 lines>...
)
datasets.exceptions.DatasetGenerationCastError: An error occurred while generating the dataset
All the data files must have the same columns, but at some point there are 3 new columns ({'network_formula_psnr_db', 'network_formula_mse', 'normalized_negative_log_measurement_density'}) and 6 missing columns ({'task', 'network_first_call_seconds', 'psnr_network_vs_formula_db', 'psnr_network_vs_ground_truth_db', 'psnr_formula_vs_ground_truth_db', 'formula_seconds'}).
This happened while the csv dataset builder was generating data using
hf://datasets/StochasticRhapsody/analytic-mmse-repro-results/measurement_density_probe.csv (at revision 5a9a868afcb6b15c39fc16d61606e9b5bb4c72c1), ['hf://datasets/StochasticRhapsody/analytic-mmse-repro-results@5a9a868afcb6b15c39fc16d61606e9b5bb4c72c1/checkpoint_formula_grid.csv', 'hf://datasets/StochasticRhapsody/analytic-mmse-repro-results@5a9a868afcb6b15c39fc16d61606e9b5bb4c72c1/measurement_density_probe.csv']
Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)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.
task string | sigma float64 | test_index int64 | psnr_network_vs_formula_db float64 | psnr_formula_vs_ground_truth_db float64 | psnr_network_vs_ground_truth_db float64 | formula_seconds float64 | network_first_call_seconds float64 | mean_log_patch_mass_unnormalized float64 |
|---|---|---|---|---|---|---|---|---|
denoising | 0.05 | 0 | 30.875734 | 29.9962 | 33.045657 | 0.517441 | 0.15471 | -7.315141 |
denoising | 0.2 | 0 | 33.674598 | 24.254593 | 23.721505 | 0.308977 | 0.005091 | 0.202671 |
denoising | 0.8 | 0 | 27.660447 | 15.003979 | 15.024639 | 0.308977 | 0.005267 | 1.737714 |
inpainting_center_15 | 0.05 | 0 | 27.708951 | 13.334259 | 13.465044 | 0.311425 | 0.005093 | -2.532987 |
inpainting_center_15 | 0.2 | 0 | 30.787519 | 13.177167 | 13.15694 | 0.308967 | 0.004882 | 1.032208 |
inpainting_center_15 | 0.8 | 0 | 31.567365 | 11.8657 | 11.845446 | 0.30903 | 0.004654 | 1.925531 |
convolution_gaussian_1.0 | 0.05 | 0 | 30.697682 | 27.466084 | 26.431607 | 0.309244 | 0.00519 | -3.269488 |
convolution_gaussian_1.0 | 0.2 | 0 | 32.434341 | 21.741525 | 21.214041 | 0.309085 | 0.004633 | -0.380209 |
convolution_gaussian_1.0 | 0.8 | 0 | 28.775094 | 14.05939 | 13.931616 | 0.309115 | 0.005136 | 2.911349 |
null | 0.05 | 0 | null | null | null | null | null | -6.950802 |
null | 0.05 | 1 | null | null | null | null | null | -32.573006 |
null | 0.05 | 2 | null | null | null | null | null | -6.218126 |
null | 0.05 | 3 | null | null | null | null | null | -10.032766 |
null | 0.05 | 4 | null | null | null | null | null | -16.969915 |
null | 0.05 | 5 | null | null | null | null | null | -8.614608 |
ICML 2026 #5221 reproduction artifacts
Raw A100 results for the reproduction of An analytic theory of convolutional neural network inverse problems solvers.
- Exact code snapshot: GitHub commit e800d23d179b163e2e5f8114b48cf1d0e7a892ae
- Authors' checkpoint repository: mhnguyen712/analytical_mmse
- Source dataset: FashionMNIST
The grid is deliberately scaled: one fixed test image for each of three tasks and three noise levels, plus six low-noise density probes. See summary.json for scope and environment, and the CSV files for raw condition-level results.
- Downloads last month
- 11