Datasets:

fryabcfry
/

severson-2019-processed

Error code:   StreamingRowsError
Exception:    CastError
Message:      Couldn't cast
id: string
category: string
status: string
run_id: string
checkpoint_available: bool
notes: list<item: string>
  child 0, item: string
bytes: int64
zip_sha256: string
metadata: struct<category: string, checkpoint_available: bool, id: string, notes: list<item: string>, run_id:  (... 23 chars omitted)
  child 0, category: string
  child 1, checkpoint_available: bool
  child 2, id: string
  child 3, notes: list<item: string>
      child 0, item: string
  child 4, run_id: string
  child 5, status: string
zip: string
file_count: int64
recipe_sha256: string
verified: bool
to
{'bytes': Value('int64'), 'category': Value('string'), 'file_count': Value('int64'), 'id': Value('string'), 'metadata': {'category': Value('string'), 'checkpoint_available': Value('bool'), 'id': Value('string'), 'notes': List(Value('string')), 'run_id': Value('string'), 'status': Value('string')}, 'recipe_sha256': Value('string'), 'status': Value('string'), 'verified': Value('bool'), 'zip': Value('string'), 'zip_sha256': Value('string')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 149, 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 129, 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 544, 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 3052, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2559, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2584, 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 558, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 440, 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 361, 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 2381, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2309, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              id: string
              category: string
              status: string
              run_id: string
              checkpoint_available: bool
              notes: list<item: string>
                child 0, item: string
              bytes: int64
              zip_sha256: string
              metadata: struct<category: string, checkpoint_available: bool, id: string, notes: list<item: string>, run_id:  (... 23 chars omitted)
                child 0, category: string
                child 1, checkpoint_available: bool
                child 2, id: string
                child 3, notes: list<item: string>
                    child 0, item: string
                child 4, run_id: string
                child 5, status: string
              zip: string
              file_count: int64
              recipe_sha256: string
              verified: bool
              to
              {'bytes': Value('int64'), 'category': Value('string'), 'file_count': Value('int64'), 'id': Value('string'), 'metadata': {'category': Value('string'), 'checkpoint_available': Value('bool'), 'id': Value('string'), 'notes': List(Value('string')), 'run_id': Value('string'), 'status': Value('string')}, 'recipe_sha256': Value('string'), 'status': Value('string'), 'verified': Value('bool'), 'zip': Value('string'), 'zip_sha256': 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.

Processed Severson battery SoH data and experiment artifacts

This repository contains a processed Severson-derived battery corpus and separately versioned artifacts for rollout-aligned probabilistic forecasting. The archives are research artifacts, not a ready-made load_dataset() table.

Existing processed data

Retain severson_processed_20260807.tar.gz. Its audited source revision is 79f40996bf8eec2fb8a3bda391ed3072d853dfd5; its size is 3,587,570,430 bytes and SHA-256 is f45812e423ba81acb3a53b5859ef3e9c1290358eb959c01d0144987159d65e58. The experiment-server copy was verified against the Hub LFS digest.

Download the pinned processed archive.

The tar archive has a data/severson-2019/processed/ root. The processed representation includes cycle arrays, cells.jsonl, cycles.jsonl, split files and train-fitted normalization metadata where present. The metadata supplement ZIP contains the exact split and normalizer files used by the release. The 140-cell processed corpus contains 114,314 cycle records; it is not claimed to be the exact original paper's 124-cell subset.

The release split is cell-level: 98 training, 21 validation and 21 test cells. With history 30 and forecast horizon 80, there are 17,317 test windows from 20 cells; 15,737 windows have a complete 80-step future. Partial future targets are masked. SoH is discharge capacity divided by the first positive finite discharge capacity of the same cell.

Array files retain their stored representation. The provided encoder configurations use input resampling to 1,000 points; do not infer model input length from an array file's stored shape. Latent caches have shape (114314, 256) and are specific to an encoder checkpoint.

Additional release files

code/: source snapshot corresponding to the linked GitHub release branch.
metadata/: split, normalization and processed-table supplement.
latents/: frozen encoder caches, separated by encoder identity.
experiments/: complete controlled runs with best/last checkpoints and available evaluations.
legacy/: completed historical training settings; not matched controls for the main model.
evaluation_only/: predictions whose original training checkpoint was not recovered.
inventory.json: original runs, snapshot identities and explicit exclusions.
release_index.json and SHA256SUMS: available archive files, sizes and hashes.

Only ZIPs listed as verified in the release index are ready for use. Pending experiments are added after the training and evaluation pipelines finish. Each ZIP includes ARTIFACT.json and a per-file FILES.json checksum manifest. Training-state snapshots that did not complete the planned budget are not silently labeled as complete runs.

The original fast-encoder TF-CRPS run (095129) has saved evaluations but no recovered training checkpoint; its archive is explicitly evaluation-only. The completed encoder8161 snapshot must be distinguished from an epoch-2 snapshot with the same directory name. The old hidden256 fair experiment is a legacy configuration, not the new controlled hidden512 fair experiment.

Code and reproduction

Code release branch (commit f9e49922496531a518d565b8bc87b74a8f21b92d).

Verify ZIP checksums before extraction. The code release provides code/scripts/package_release.py extract, which validates member hashes and refuses unsafe paths or overwriting an existing directory. Checkpoint-embedded paths refer to the source machines; use the portable config where supplied or explicit --latent-cache, --split-file and --output-dir overrides.

The encoder42 frozen cache supports predictor reproduction. The matching original encoder42 training checkpoint has not been recovered, so this release does not claim complete raw-curve-to-output reproduction for that encoder.

Interpretation

One sampled VBLL output-layer weight is held fixed along each forecast trajectory. No additional observation noise is sampled. Intervals therefore represent weight-posterior-induced trajectory variation. All probabilistic comparison metrics use empirical CRPS with 128 evaluation samples, including models trained with fair CRPS. Keep micro/macro averaging and best/last checkpoint results distinct.

Attribution and terms

This is derived processed data. Preserve the original dataset's attribution and applicable terms; the artifact release does not replace them. The existing repository's other license designation is retained, not replaced by a newly asserted permissive license.

Severson, K. A., Attia, P. M., Jin, N., et al. Data-driven prediction of battery cycle life before capacity degradation. Nature Energy 4, 383-391 (2019). https://doi.org/10.1038/s41560-019-0356-8

Downloads last month: 34

Total file size:

20.6 GB