The dataset viewer is not available for this split.
Error code: StreamingRowsError
Exception: ArrowInvalid
Message: JSON parse error: Invalid value. in row 0
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 342, in _generate_tables
df = pandas_read_json(f)
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 38, in pandas_read_json
return pd.read_json(path_or_buf, **kwargs)
~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/pandas/io/json/_json.py", line 815, in read_json
return json_reader.read()
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/pandas/io/json/_json.py", line 1014, in read
obj = self._get_object_parser(self.data)
File "/usr/local/lib/python3.14/site-packages/pandas/io/json/_json.py", line 1040, in _get_object_parser
obj = FrameParser(json, **kwargs).parse()
File "/usr/local/lib/python3.14/site-packages/pandas/io/json/_json.py", line 1176, in parse
self._parse()
~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/pandas/io/json/_json.py", line 1392, in _parse
ujson_loads(json, precise_float=self.precise_float), dtype=None
~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
ValueError: Expected object or value
During handling of the above exception, another exception occurred:
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 345, in _generate_tables
raise e
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 291, in _generate_tables
pa_table = paj.read_json(
io.BytesIO(batch), read_options=paj.ReadOptions(block_size=block_size)
)
File "pyarrow/_json.pyx", line 364, in pyarrow._json.read_json
File "pyarrow/error.pxi", line 155, in pyarrow.lib.pyarrow_internal_check_status
File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
raise convert_status(status)
pyarrow.lib.ArrowInvalid: JSON parse error: Invalid value. in row 0Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
Precious Cargo — Reroll Rumble test record
Three independent versions of one six-clip story, made with MiniMax H3, Singularity v1.3 and Singularity V2. This archive contains all33 delivered scored outputs shown in the episode, including rejected rolls. Public availability and anonymous remote archive verification are recorded separately in the release receipts.
What was tested
Native ComfyUI H3 image-to-video: shared starting image per scene, plus shared ending images for scenes2 and5;1344×768,24fps,20steps,res_multistep,simple schedule,BasicGuider, no acceleration adapter. Full workflows, exact conditioning strings, seeds, input images, model hashes and per-run receipt exports are included. Original frame counts by scene:294,243,192,243,243,294 (62.875seconds). Generated scene audio is preserved in originals. All33 complete original MP4s and33 API workflows retain their saved SHA-256 values.
Singularity v1.3 used10 scored rolls (4rerolls); MiniMax H3 used16 (10rerolls); Singularity V2 used7 (1reroll). Each has6 individual keepers. First roll plus up to7rerolls per model/scene was the quality-attempt limit. Provider attempt IDs can differ because technical recoveries are not quality rerolls. Historic café B/C labels differed; the final continuing-film mapping is recorded explicitly. Current episode uses A=v1.3,B=H3,C=V2 before its reveal. Raw history and corrections remain separate from the normalized final decisions.
Costs and limits
Estimated GPU-job spend, including rejected scored outputs: Singularity v1.3 $7.07, MiniMax H3 $12.97, Singularity V2 $4.89; combined unrounded sum $24.92. Means per individual keeper: $1.18,$2.16,$0.81. Exact formula: provider executionTime milliseconds /1000 /3600 × historical $3.49/GPU-hour ×1GPU, RTX PRO6000 Blackwell Server Edition96GB. Processing/setup/downloads inside a job count. Queue time, outside-job startup/idle, storage, unscored work, local editing and audio review do not. No per-job ceiling was invented. Sum full precision before rounding display cents. Generation-call timing is an alternate recorded measure, not the chosen main cost basis or isolated GPU-kernel time.
About$32 in posted identified-resource charges is a separate overlapping view including preflight and technical recoveries. Storage coverage is incomplete; this is not a final invoice. Never add it to the scored-job estimates or subtract the two to claim exact overhead. Supplemental dedicated-endpoint-session charges are not pure per-clip invoices. See costs.json for inputs and provenance; actual per-clip invoices remain null where unavailable.
Keepers satisfy the original declared action/continuity criteria. The creator explicitly found the films unconvincing in realism; passing is not an appearance-quality endorsement. Future Rumble realism/physical-plausibility gates were not applied retroactively. Adjacent cuts/assembled-film acceptance remain separate from individual-clip decisions. A4's accepted editorial tail trim retains frames0–208 of243; originals are complete. Selected films are presentation derivatives with leveled native sound. C3 has disclosed selective speech cleanup; originals retain raw audio. No generic added ambience.
Reuse and archive scope
Use scored-clips.json to resolve each video, exact workflow, matching input files, dependencies, receipt export, final judgment and cost. Copy reference images into ComfyUI input using each entry's comfy_input_name; inputs are deduplicated by hash. Inference-code files are preserved execution-source snapshots, not a turnkey spending controller: supply your own appropriately authorized environment and dependencies. Checkpoint-lock contains historical proposal metadata; actual per-run model_proof hashes are the downloaded-byte evidence. Third-party code/model licenses remain with their respective publishers; model weights are not redistributed here.
Includes33 scored originals,33 exact workflows, all input references used by those graphs, selected receipt fields/dependency freezes, saved creator-review exports, final judgments/trim/audio policy, cost ledger, selected-film derivatives and graphics/inference source snapshots. Excludes model weights, original voice-reference audio, generated host-voice WAVs, full private episode masters, credentials, provider session URLs, deployment bundles and unrelated account data. Unscored smoke tests, failed transport fragments and full operational logs are not reproduced; attributable costs and technical-recovery scope are retained in the ledger. This is not a claim to contain every provider job or every private production file. JSON review/receipt exports normalize local file paths and identify their original hashes; they are not verbatim raw receipts. The historical label-map note is evidence bookkeeping, not new viewer narration.
The episode's4K delivery is an upscale of its1080p composition, not native4K model output. Caption/narration text is provided; no voice reference is public. Consult the separate release receipts for creator approval, public repository/archive verification and publication state. These source records do not authorize new spending or publication.
Download
SHA-256: d1312abfb8dd20de9acbd021bb1c799dd3a3d25118a0682a007e769e2a2b82d7
The ZIP contains its per-file manifest. Scored clip index, per-roll costs and source links are available separately.
- Downloads last month
- 47