The dataset viewer is not available for this split.
Error code: StreamingRowsError
Exception: CastError
Message: Couldn't cast
schema_version: int64
name: string
verified_at: timestamp[s]
source_manifest_sha256: string
config_sha256: string
ordering: struct<chunk_separator: string, document_separator_hex: string, group_interleave: string, shuffle_bu (... 81 chars omitted)
child 0, chunk_separator: string
child 1, document_separator_hex: string
child 2, group_interleave: string
child 3, shuffle_buffer_documents: int64
child 4, shuffle_unit: string
child 5, within_document_chunk_order: string
redistribution: struct<public_raw_data_allowed: bool, reason: string>
child 0, public_raw_data_allowed: bool
child 1, reason: string
comparison_contract: struct<same_source_bytes: bool, same_order_and_offsets: bool, same_reset_policy: bool, same_curricul (... 107 chars omitted)
child 0, same_source_bytes: bool
child 1, same_order_and_offsets: bool
child 2, same_reset_policy: bool
child 3, same_curriculum_and_quality_filters: bool
child 4, same_validation_sha256: bool
child 5, data_changes_require_separate_experiment: bool
splits: struct<train: struct<bytes: int64, sha256: string, matches_historical_manifest: bool>, validation: s (... 71 chars omitted)
child 0, train: struct<bytes: int64, sha256: string, matches_historical_manifest: bool>
child 0, bytes: int64
child 1, sha256: string
child 2, matches_historical_manifest: bool
child 1, validation: struct<bytes: int64, sha256: string, matches_historical_manifest: bool>
child 0, bytes: int64
child 1, sha256: string
child 2, matches_historical_manifest: bool
bytes: int64
matches_historical_manifest: bool
corpus: string
split: string
sha256: string
to
{'corpus': Value('string'), 'split': Value('string'), 'bytes': Value('int64'), 'sha256': Value('string'), 'matches_historical_manifest': Value('bool')}
because column names don't match
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 2951, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2461, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2486, 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 547, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 430, 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 343, 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 2378, in table_cast
return cast_table_to_schema(table, schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
raise CastError(
...<3 lines>...
)
datasets.table.CastError: Couldn't cast
schema_version: int64
name: string
verified_at: timestamp[s]
source_manifest_sha256: string
config_sha256: string
ordering: struct<chunk_separator: string, document_separator_hex: string, group_interleave: string, shuffle_bu (... 81 chars omitted)
child 0, chunk_separator: string
child 1, document_separator_hex: string
child 2, group_interleave: string
child 3, shuffle_buffer_documents: int64
child 4, shuffle_unit: string
child 5, within_document_chunk_order: string
redistribution: struct<public_raw_data_allowed: bool, reason: string>
child 0, public_raw_data_allowed: bool
child 1, reason: string
comparison_contract: struct<same_source_bytes: bool, same_order_and_offsets: bool, same_reset_policy: bool, same_curricul (... 107 chars omitted)
child 0, same_source_bytes: bool
child 1, same_order_and_offsets: bool
child 2, same_reset_policy: bool
child 3, same_curriculum_and_quality_filters: bool
child 4, same_validation_sha256: bool
child 5, data_changes_require_separate_experiment: bool
splits: struct<train: struct<bytes: int64, sha256: string, matches_historical_manifest: bool>, validation: s (... 71 chars omitted)
child 0, train: struct<bytes: int64, sha256: string, matches_historical_manifest: bool>
child 0, bytes: int64
child 1, sha256: string
child 2, matches_historical_manifest: bool
child 1, validation: struct<bytes: int64, sha256: string, matches_historical_manifest: bool>
child 0, bytes: int64
child 1, sha256: string
child 2, matches_historical_manifest: bool
bytes: int64
matches_historical_manifest: bool
corpus: string
split: string
sha256: string
to
{'corpus': Value('string'), 'split': Value('string'), 'bytes': Value('int64'), 'sha256': Value('string'), 'matches_historical_manifest': Value('bool')}
because column names don't matchNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
Monumental Models data manifests
Content-addressed dataset identity records for the Monumental Models research program. This repository contains metadata only, not the training text, teacher outputs or model checkpoints.
The recovered document-corpus-v3 training and validation streams were re-hashed on October12,2026 and match the original July28 build manifest. data/splits.jsonl contains one manifest record per split; its two rows are not the number of training documents. The original document index contains2,508,634records.
Use identical source hashes, ordering, curriculum, split/reset policy and consumed-byte budgets for controlled architecture comparisons. Altering data is a separately registered experiment. Counts here describe available corpus bytes, not training exposure or token counts.
The source mixture includes curated material and NVIDIA collections. Public raw redistribution has not been approved by this audit. Access grants and this public metadata repository do not license the underlying text. Retain source-specific pinned terms and publish only eligible artifacts. The release decision is recorded separately from content identity. No raw samples or machine-local paths are included.
The quality pipeline is MonumentalSystems text-pipeline followed by NeMo Curator. Some trimmed exports dropped readability/curriculum fields; those labels must be recovered or recomputed separately without modifying the reference text. Manifest hashes allow comparison without republishing restricted data.
- Downloads last month
- 35