Datasets:

Srishti280992
/

repro-well-posed-kl-control-bundle

Modalities:

Dataset Preview

Split (1)

train

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 19 new columns ({'KW_beats_KL_rate', 'KW_max_abs_x_mean', 'WKL_max_abs_x_mean', 'KL_over_KW_state_cost', 'KL_final_norm_mean', 'KW_control_cost_mean', 'KL_max_abs_phi_mean', 'WKL_state_cost_mean', 'KW_max_abs_phi_mean', 'KL_control_cost_mean', 'KL_max_abs_x_mean', 'KL_state_cost_mean', 'KW_state_cost_mean', 'WKL_final_norm_mean', 'WKL_control_cost_mean', 'WKL_beats_KL_rate', 'KW_final_norm_mean', 'WKL_max_abs_phi_mean', 'KL_over_WKL_state_cost'}) and 7 missing columns ({'final_norm', 'seed', 'max_abs_x', 'max_abs_phi', 'state_cost', 'controller', 'control_cost'}).

This happened while the csv dataset builder was generating data using

hf://datasets/Srishti280992/repro-well-posed-kl-control-bundle/outputs/local_repro/cartpole_summary.csv (at revision a316609daaf334d4869bcfcb5e417fdc75a4e937), ['hf://datasets/Srishti280992/repro-well-posed-kl-control-bundle@a316609daaf334d4869bcfcb5e417fdc75a4e937/outputs/local_repro/cartpole_runs.csv', 'hf://datasets/Srishti280992/repro-well-posed-kl-control-bundle@a316609daaf334d4869bcfcb5e417fdc75a4e937/outputs/local_repro/cartpole_summary.csv', 'hf://datasets/Srishti280992/repro-well-posed-kl-control-bundle@a316609daaf334d4869bcfcb5e417fdc75a4e937/outputs/local_repro/double_integrator_mc.csv', 'hf://datasets/Srishti280992/repro-well-posed-kl-control-bundle@a316609daaf334d4869bcfcb5e417fdc75a4e937/outputs/local_repro/double_integrator_sweep.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
              rho: double
              KL_state_cost_mean: double
              KL_control_cost_mean: double
              KL_max_abs_x_mean: double
              KL_max_abs_phi_mean: double
              KL_final_norm_mean: double
              WKL_state_cost_mean: double
              WKL_control_cost_mean: double
              WKL_max_abs_x_mean: double
              WKL_max_abs_phi_mean: double
              WKL_final_norm_mean: double
              KW_state_cost_mean: double
              KW_control_cost_mean: double
              KW_max_abs_x_mean: double
              KW_max_abs_phi_mean: double
              KW_final_norm_mean: double
              WKL_beats_KL_rate: double
              KL_over_WKL_state_cost: double
              KW_beats_KL_rate: double
              KL_over_KW_state_cost: double
              -- schema metadata --
              pandas: '{"index_columns": [{"kind": "range", "name": null, "start": 0, "' + 3017
              to
              {'rho': Value('float64'), 'seed': Value('int64'), 'controller': Value('string'), 'state_cost': Value('float64'), 'control_cost': Value('float64'), 'max_abs_x': Value('float64'), 'max_abs_phi': Value('float64'), 'final_norm': 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 19 new columns ({'KW_beats_KL_rate', 'KW_max_abs_x_mean', 'WKL_max_abs_x_mean', 'KL_over_KW_state_cost', 'KL_final_norm_mean', 'KW_control_cost_mean', 'KL_max_abs_phi_mean', 'WKL_state_cost_mean', 'KW_max_abs_phi_mean', 'KL_control_cost_mean', 'KL_max_abs_x_mean', 'KL_state_cost_mean', 'KW_state_cost_mean', 'WKL_final_norm_mean', 'WKL_control_cost_mean', 'WKL_beats_KL_rate', 'KW_final_norm_mean', 'WKL_max_abs_phi_mean', 'KL_over_WKL_state_cost'}) and 7 missing columns ({'final_norm', 'seed', 'max_abs_x', 'max_abs_phi', 'state_cost', 'controller', 'control_cost'}).
              
              This happened while the csv dataset builder was generating data using
              
              hf://datasets/Srishti280992/repro-well-posed-kl-control-bundle/outputs/local_repro/cartpole_summary.csv (at revision a316609daaf334d4869bcfcb5e417fdc75a4e937), ['hf://datasets/Srishti280992/repro-well-posed-kl-control-bundle@a316609daaf334d4869bcfcb5e417fdc75a4e937/outputs/local_repro/cartpole_runs.csv', 'hf://datasets/Srishti280992/repro-well-posed-kl-control-bundle@a316609daaf334d4869bcfcb5e417fdc75a4e937/outputs/local_repro/cartpole_summary.csv', 'hf://datasets/Srishti280992/repro-well-posed-kl-control-bundle@a316609daaf334d4869bcfcb5e417fdc75a4e937/outputs/local_repro/double_integrator_mc.csv', 'hf://datasets/Srishti280992/repro-well-posed-kl-control-bundle@a316609daaf334d4869bcfcb5e417fdc75a4e937/outputs/local_repro/double_integrator_sweep.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.

rho float64	seed int64	controller string	state_cost float64	control_cost float64	max_abs_x float64	max_abs_phi float64	final_norm float64
0.001	0	KL	1.859737	3.456581	4.282011	0.605874	0.77726
0.001	0	WKL	0.566415	18.225804	1.671878	0.361603	0.097944
0.001	0	KW	0.566407	18.231888	1.671856	0.3616	0.097936
0.001	1	KL	2.003236	3.988626	4.240966	0.682946	2.913915
0.001	1	WKL	0.630714	18.3502	1.859114	0.457927	0.567417
0.001	1	KW	0.630705	18.356058	1.859094	0.457923	0.567415
0.001	2	KL	2.942065	4.47487	5.040887	0.590931	3.850978
0.001	2	WKL	0.811986	18.728303	2.253016	0.478246	1.852935
0.001	2	KW	0.811971	18.734288	2.25299	0.478238	1.852933
0.001	3	KL	2.135298	3.639932	4.203137	0.641867	1.856612
0.001	3	WKL	0.586373	18.664412	1.973585	0.504163	1.006693
0.001	3	KW	0.586364	18.670637	1.973565	0.504156	1.006681
0.001	4	KL	2.428793	4.309765	4.96315	0.764902	2.710467
0.001	4	WKL	0.739253	18.842106	2.390625	0.585139	0.998471
0.001	4	KW	0.739241	18.848048	2.3906	0.585129	0.998452
0.001	5	KL	2.227478	4.037119	4.979252	0.61879	1.169071
0.001	5	WKL	0.702344	19.008999	2.130596	0.428524	0.67676
0.001	5	KW	0.702334	19.014926	2.130569	0.428516	0.676764
0.001	6	KL	2.256308	4.166567	4.802728	0.690986	4.264059
0.001	6	WKL	0.669488	18.270834	1.728499	0.460745	1.390165
0.001	6	KW	0.669477	18.276744	1.728473	0.460737	1.390152
0.001	7	KL	2.917223	4.090818	4.864657	0.637837	3.426875
0.001	7	WKL	0.765076	18.610686	2.099709	0.435969	2.148272
0.001	7	KW	0.765062	18.616741	2.099685	0.43596	2.148296
0.001	8	KL	2.896839	3.875577	4.63017	0.57724	3.284667
0.001	8	WKL	0.771842	18.598861	1.965665	0.449715	0.81052
0.001	8	KW	0.771828	18.604749	1.965638	0.449706	0.810519
0.001	9	KL	2.352566	4.131793	5.32924	0.814357	1.637891
0.001	9	WKL	0.689551	18.922206	1.952437	0.449713	0.779096
0.001	9	KW	0.689539	18.928247	1.952403	0.449705	0.779086
0.001	10	KL	3.125845	4.528431	4.809757	0.57733	1.503007
0.001	10	WKL	0.865446	18.295968	2.332288	0.486673	0.930133
0.001	10	KW	0.86543	18.301794	2.332271	0.486665	0.930144
0.001	11	KL	3.019433	4.566026	4.791571	0.648583	2.407002
0.001	11	WKL	0.8348	18.464114	2.168933	0.460801	1.31742
0.001	11	KW	0.834785	18.469881	2.168908	0.460795	1.317419
0.001	12	KL	2.188282	3.582807	4.808413	0.639744	0.48288
0.001	12	WKL	0.60358	18.417429	1.939383	0.499047	1.357582
0.001	12	KW	0.60357	18.423527	1.939364	0.499042	1.357598
0.001	13	KL	2.665508	3.999414	5.748886	0.554874	1.653783
0.001	13	WKL	0.734363	18.633275	2.150609	0.463471	0.869412
0.001	13	KW	0.73435	18.639355	2.150586	0.463463	0.869409
0.001	14	KL	2.102885	4.175948	3.849405	0.626422	1.189073
0.001	14	WKL	0.669255	18.425555	1.821013	0.470366	0.269538
0.001	14	KW	0.669244	18.431483	1.820997	0.470362	0.269529
0.001	15	KL	2.08916	4.072854	5.177279	0.779235	0.884422
0.001	15	WKL	0.64499	18.667788	1.938913	0.453755	0.922279
0.001	15	KW	0.644979	18.673751	1.938889	0.453746	0.92222
0.001	16	KL	2.391264	4.433005	4.792643	0.722762	4.105141
0.001	16	WKL	0.766214	18.699348	2.031991	0.442997	1.94358
0.001	16	KW	0.766201	18.705313	2.031967	0.442988	1.943586
0.001	17	KL	3.036426	4.363701	6.994498	0.775755	1.071693
0.001	17	WKL	0.779898	18.664481	2.843799	0.591251	0.49491
0.001	17	KW	0.779884	18.67049	2.843761	0.59124	0.494895
0.001	18	KL	3.102641	3.747648	5.574125	0.763935	1.594047
0.001	18	WKL	0.726387	18.580983	2.307617	0.480569	0.714566
0.001	18	KW	0.726375	18.587007	2.307586	0.48056	0.71457
0.001	19	KL	2.070832	3.886918	4.065987	0.572973	1.770313
0.001	19	WKL	0.623239	18.636443	1.774103	0.462324	0.347025
0.001	19	KW	0.623229	18.642488	1.77408	0.462314	0.347043
0.001	20	KL	3.094506	4.334439	4.603671	0.63223	1.089499
0.001	20	WKL	0.75914	18.859906	2.151954	0.46921	0.478004
0.001	20	KW	0.759126	18.866077	2.151931	0.469203	0.477997
0.001	21	KL	1.933453	3.995966	4.136484	0.632395	2.844398
0.001	21	WKL	0.665917	18.561672	1.903309	0.415328	2.055924
0.001	21	KW	0.665907	18.567708	1.90329	0.415321	2.055889
0.001	22	KL	3.142895	4.638976	4.861155	0.769349	2.39947
0.001	22	WKL	0.786224	19.156993	1.721587	0.472089	1.191396
0.001	22	KW	0.786209	19.163012	1.721567	0.472082	1.191313
0.001	23	KL	3.625436	4.133914	6.295848	0.726483	4.038235
0.001	23	WKL	0.816276	18.430478	2.542343	0.516228	1.382184
0.001	23	KW	0.81626	18.436438	2.542309	0.516218	1.382143
0.001	24	KL	2.426778	4.154098	4.651379	0.79978	0.888776
0.001	24	WKL	0.695478	18.598254	2.133036	0.511649	0.252741
0.001	24	KW	0.695466	18.604236	2.133009	0.511638	0.252704
0.001	25	KL	2.232849	4.192015	4.287785	0.605461	4.272459
0.001	25	WKL	0.695624	17.914768	1.844349	0.554173	3.062752
0.001	25	KW	0.695612	17.920591	1.844335	0.554163	3.0627
0.001	26	KL	3.463091	4.479254	5.803169	0.663418	4.493114
0.001	26	WKL	0.872994	18.579738	2.301547	0.462548	2.245089
0.001	26	KW	0.872978	18.585647	2.301519	0.462538	2.245065
0.001	27	KL	1.867289	4.290818	4.883287	0.725298	1.759954
0.001	27	WKL	0.676829	18.414487	2.604526	0.447328	1.741264
0.001	27	KW	0.676819	18.420251	2.604496	0.447323	1.741206
0.001	28	KL	3.203916	4.111105	5.28828	0.630271	3.37429
0.001	28	WKL	0.920299	18.857188	2.880158	0.448907	1.021194
0.001	28	KW	0.920285	18.863117	2.880146	0.448898	1.021179
0.001	29	KL	2.693515	4.107427	5.682969	0.592846	1.025633
0.001	29	WKL	0.745951	18.226941	2.108319	0.440916	0.471034
0.001	29	KW	0.745939	18.23284	2.108291	0.440908	0.471016
0.001	30	KL	2.776996	3.734683	5.576316	0.617035	0.759497
0.001	30	WKL	0.704888	18.619703	1.941077	0.448106	1.319426
0.001	30	KW	0.704876	18.625882	1.941045	0.448102	1.319413
0.001	31	KL	2.784331	4.247334	4.751623	0.693533	3.098241
0.001	31	WKL	0.718572	18.491272	2.100993	0.549232	1.61398
0.001	31	KW	0.718558	18.49715	2.100967	0.549225	1.613986
0.0001	0	KL	1.48656	0.265351	4.290403	0.17868	2.032798
0.0001	0	WKL	0.056605	1.81137	0.529874	0.114909	0.028695
0.0001	0	KW	0.056605	1.811431	0.529873	0.114909	0.028694
0.0001	1	KL	1.570495	0.297874	5.121406	0.186647	3.381555

End of preview.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

Reproduction Bundle

Paper: Well-Posed KL-Regularized Control via Wasserstein and Kalman-Wasserstein KL Divergences

This bundle contains the independent NumPy/SciPy reproduction used in the Trackio logbook:

scripts/reproduce_kwkl_control.py: implementation of Eq. (15), Eq. (16), the double-integrator sweep, anisotropic double-integrator check, and nonlinear cart-pole simulation.
outputs/local_repro/summary.json: final local 32-seed cart-pole and 64-seed double-integrator replay summary.
outputs/local_repro/*.csv: raw tables for the final local run.
outputs/local_repro/*.html: simple rendered result tables used as logbook figures.
sources/paper.pdf and sources/paper_text_clean.txt: paper copy and extracted text used to map equations/sections.

Rerun from the logbook workspace with:

/Users/srishtisaha/Desktop/icml-2026-agent-repro/.venv/bin/python scripts/reproduce_kwkl_control.py --outdir outputs/local_repro --cartpole-seeds 32 --linear-seeds 64

Downloads last month: 31

Total file size:

1.01 MB