The dataset viewer is not available for this split.
Error code: StreamingRowsError
Exception: CastError
Message: Couldn't cast
id: string
title: string
world: string
prompt: string
prompt_is: string
observation: string
seed: int64
lora: string
trigger: string
film: struct<path: string, duration: double, bytes: int64, sha256: string>
child 0, path: string
child 1, duration: double
child 2, bytes: int64
child 3, sha256: string
provenance: list<item: struct<world: string, model_config: string, model_revision: string, config_sha256: string (... 74 chars omitted)
child 0, item: struct<world: string, model_config: string, model_revision: string, config_sha256: string, source_da (... 62 chars omitted)
child 0, world: string
child 1, model_config: string
child 2, model_revision: string
child 3, config_sha256: string
child 4, source_dataset: string
child 5, identity_evidence: string
child 6, limitations: string
clips: list<item: struct<id: string, order: int64, title: string, world: string, prompt: string, prompt_is: (... 280 chars omitted)
child 0, item: struct<id: string, order: int64, title: string, world: string, prompt: string, prompt_is: string, ob (... 268 chars omitted)
child 0, id: string
child 1, order: int64
child 2, title: string
child 3, world: string
child 4, prompt: string
child 5, prompt_is: string
child 6, observation: string
child 7, seed: int64
child 8, lora: string
child 9, trigger: string
child 10, path: string
child 11, duration: double
child 12, start: double
child 13, width: int64
child 14, height: int64
child 15, frames: int64
child 16, fps: string
child 17, audio_streams: int64
child 18, bytes: int64
child 19, sha256: string
child 20, raw_source_sha256: string
child 21, privacy_remux: string
creator: string
contract: struct<base: string, VDN: string, turbo: bool, sampler: string, scheduler: string, steps: int64, meg (... 148 chars omitted)
child 0, base: string
child 1, VDN: string
child 2, turbo: bool
child 3, sampler: string
child 4, scheduler: string
child 5, steps: int64
child 6, megapixels: double
child 7, fps: int64
child 8, requested_seconds: int64
child 9, actual_frames_per_scene: int64
child 10, scene_lora_strength: int64
child 11, only_changed: list<item: string>
child 0, item: string
assembly: string
original_workflow_sha256: string
disclosure: string
to
{'title': Value('string'), 'creator': Value('string'), 'disclosure': Value('string'), 'assembly': Value('string'), 'original_workflow_sha256': Value('string'), 'contract': {'base': Value('string'), 'VDN': Value('string'), 'turbo': Value('bool'), 'sampler': Value('string'), 'scheduler': Value('string'), 'steps': Value('int64'), 'megapixels': Value('float64'), 'fps': Value('int64'), 'requested_seconds': Value('int64'), 'actual_frames_per_scene': Value('int64'), 'scene_lora_strength': Value('int64'), 'only_changed': List(Value('string'))}, 'provenance': List({'world': Value('string'), 'model_config': Value('string'), 'model_revision': Value('string'), 'config_sha256': Value('string'), 'source_dataset': Value('string'), 'identity_evidence': Value('string'), 'limitations': Value('string')}), 'clips': List({'id': Value('string'), 'order': Value('int64'), 'title': Value('string'), 'world': Value('string'), 'prompt': Value('string'), 'prompt_is': Value('string'), 'observation': Value('string'), 'seed': Value('int64'), 'lora': Value('string'), 'trigger': Value('string'), 'path': Value('string'), 'duration': Value('float64'), 'start': Value('float64'), 'width': Value('int64'), 'height': Value('int64'), 'frames': Value('int64'), 'fps': Value('string'), 'audio_streams': Value('int64'), 'bytes': Value('int64'), 'sha256': Value('string'), 'raw_source_sha256': Value('string'), 'privacy_remux': Value('string')}), 'film': {'path': Value('string'), 'duration': Value('float64'), 'bytes': Value('int64'), 'sha256': Value('string')}}
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 2840, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2373, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2398, 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 536, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 419, 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
id: string
title: string
world: string
prompt: string
prompt_is: string
observation: string
seed: int64
lora: string
trigger: string
film: struct<path: string, duration: double, bytes: int64, sha256: string>
child 0, path: string
child 1, duration: double
child 2, bytes: int64
child 3, sha256: string
provenance: list<item: struct<world: string, model_config: string, model_revision: string, config_sha256: string (... 74 chars omitted)
child 0, item: struct<world: string, model_config: string, model_revision: string, config_sha256: string, source_da (... 62 chars omitted)
child 0, world: string
child 1, model_config: string
child 2, model_revision: string
child 3, config_sha256: string
child 4, source_dataset: string
child 5, identity_evidence: string
child 6, limitations: string
clips: list<item: struct<id: string, order: int64, title: string, world: string, prompt: string, prompt_is: (... 280 chars omitted)
child 0, item: struct<id: string, order: int64, title: string, world: string, prompt: string, prompt_is: string, ob (... 268 chars omitted)
child 0, id: string
child 1, order: int64
child 2, title: string
child 3, world: string
child 4, prompt: string
child 5, prompt_is: string
child 6, observation: string
child 7, seed: int64
child 8, lora: string
child 9, trigger: string
child 10, path: string
child 11, duration: double
child 12, start: double
child 13, width: int64
child 14, height: int64
child 15, frames: int64
child 16, fps: string
child 17, audio_streams: int64
child 18, bytes: int64
child 19, sha256: string
child 20, raw_source_sha256: string
child 21, privacy_remux: string
creator: string
contract: struct<base: string, VDN: string, turbo: bool, sampler: string, scheduler: string, steps: int64, meg (... 148 chars omitted)
child 0, base: string
child 1, VDN: string
child 2, turbo: bool
child 3, sampler: string
child 4, scheduler: string
child 5, steps: int64
child 6, megapixels: double
child 7, fps: int64
child 8, requested_seconds: int64
child 9, actual_frames_per_scene: int64
child 10, scene_lora_strength: int64
child 11, only_changed: list<item: string>
child 0, item: string
assembly: string
original_workflow_sha256: string
disclosure: string
to
{'title': Value('string'), 'creator': Value('string'), 'disclosure': Value('string'), 'assembly': Value('string'), 'original_workflow_sha256': Value('string'), 'contract': {'base': Value('string'), 'VDN': Value('string'), 'turbo': Value('bool'), 'sampler': Value('string'), 'scheduler': Value('string'), 'steps': Value('int64'), 'megapixels': Value('float64'), 'fps': Value('int64'), 'requested_seconds': Value('int64'), 'actual_frames_per_scene': Value('int64'), 'scene_lora_strength': Value('int64'), 'only_changed': List(Value('string'))}, 'provenance': List({'world': Value('string'), 'model_config': Value('string'), 'model_revision': Value('string'), 'config_sha256': Value('string'), 'source_dataset': Value('string'), 'identity_evidence': Value('string'), 'limitations': Value('string')}), 'clips': List({'id': Value('string'), 'order': Value('int64'), 'title': Value('string'), 'world': Value('string'), 'prompt': Value('string'), 'prompt_is': Value('string'), 'observation': Value('string'), 'seed': Value('int64'), 'lora': Value('string'), 'trigger': Value('string'), 'path': Value('string'), 'duration': Value('float64'), 'start': Value('float64'), 'width': Value('int64'), 'height': Value('int64'), 'frames': Value('int64'), 'fps': Value('string'), 'audio_streams': Value('int64'), 'bytes': Value('int64'), 'sha256': Value('string'), 'raw_source_sha256': Value('string'), 'privacy_remux': Value('string')}), 'film': {'path': Value('string'), 'duration': Value('float64'), 'bytes': Value('int64'), 'sha256': Value('string')}}
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.
worldweaver — Location Seeks Cast
A Mind Expander production. Promotional fictional AI-generated video and audio.
Programme concept (not a claim of exact output adherence): A location introduces itself, demonstrates atmosphere, and auditions for a role that may be too small.
Watch/download the assembled film · Central festival screening · Manifest · Prompts and observed descriptions
Programme
3 independently rendered scenes; measured assembled duration 45.281 seconds. Intended speech may be unclear, absent, or different. Captions are visual descriptions in Markdown and optional WebVTT, explicitly not word-perfect speech transcripts.
| # | Scene | Measured duration | Model world |
|---|---|---|---|
| 1 | Location seeks cast | 15.083 s | worldweaver |
| 2 | Window department | 15.083 s | worldweaver |
| 3 | Audition for a room | 15.083 s | worldweaver |
What is in this pack
clips/: generated scenes, audiovisual streams preserved, container metadata removed for privacy.showcase-reel.mp4: straight editorial assembly preserving generated audio; cuts are intentional, not seamless I2V chaining.workflows/: privacy-safe API graphs, output prefixes generalized. No pod address or queue IDs.prompts.json,captions/: intended prompts separated from inspected visual observations.manifest.json,SHA256SUMS,posters/: measured streams, checksums and actual output frames.
Inference contract
The owner's supplied MiniMax H3 VDN graph was retained: minimax_h3_fl2va_pruned_int8_convrot.safetensors, VDN stage-dmd-step-250, turbo enabled, Euler/simple, 8 steps, 0.8 MP, widescreen, 24 fps, scene LoRA strength 1. Only node 6206 model filename, 6005 prompt, 221 duration, 6141 seed and 6170 output prefix changed. A single world LoRA is used per render; the supplied acceleration adapter is unchanged. The duration input is 15 seconds; its preserved frame-alignment formula yields 362 frames, about 15.083 seconds. Measured container duration may include small audio/container rounding.
This is a prerecorded promotional showcase, not a live inference service or speed benchmark. Character identity, intelligible dialogue and recurring theater props are not guaranteed. No extra rerenders were silently substituted.
Identity evidence and source provenance
- worldweaver: training config · source dataset reference Cinematic architectural and environmental spaces, not a documented humanoid character. Inspected examples include an empty dilapidated workshop and a bright arched church interior. Trigger worldweaver, is recorded in the model config. This showcase uses an original empty-theater setting; it does not claim a trained wizard identity. Observed outputs have pronounced panoramic/fisheye geometry and are delivered as flat videos, not a verified 360-degree experience. No source dataset README was present.
Model configurations, source captions, representative dataset thumbnails and trained preview imagery informed scene prompts. They are references only: no training/source video was used as a generated film clip. Access to referenced training datasets may require permission. Their public availability is not promised. The source datasets inspected for the four new worlds had no README, so no source-license statement could be verified there. The release does not redistribute their reference images or private source captions.
Rights and limitations
LoRAs and creative direction credited to Mind Expander as supplied by the owner. Base model: MiniMax H3. Production assistance: Hermes Agent. This credit is not a blanket assignment or clearance of all source, model, soundtrack, adapter or output rights. No new license is asserted for third-party source material. In particular, SWINGUSINNERZ is source-derived; public-domain conclusions depend on source transfer, soundtrack and jurisdiction. Review reuse for copied details, local law and applicable model terms. No claims of actual crowds, engagement, endorsements, bookings or commercial success are made.
Download
hf download TheMindExpansionNetwork/worldweaver-h3-premiere-showcase --repo-type dataset --local-dir ./worldweaver-h3-premiere-showcase
cd worldweaver-h3-premiere-showcase
sha256sum -c SHA256SUMS
For source fidelity, the untouched pod outputs remain in the private production workspace; published clips remove embedded workflow/container metadata while retaining compressed audio/video stream payloads. Hashes for both raw and released clip files are in the manifest.
- Downloads last month
- 3