Dataset Viewer
The dataset viewer is not available for this split.
Cannot load the dataset split (in streaming mode) to extract the first rows.
Error code: StreamingRowsError
Exception: CastError
Message: Couldn't cast
count: int64
unique_ids: int64
unique_sessions: int64
all_engineering_passed: bool
source_revision: string
semantic_quality: string
results: list<item: struct<id: string, batch: string, run: string, thread_id: string, query_unchanged: bool, (... 6870 chars omitted)
child 0, item: struct<id: string, batch: string, run: string, thread_id: string, query_unchanged: bool, inputs_veri (... 6858 chars omitted)
child 0, id: string
child 1, batch: string
child 2, run: string
child 3, thread_id: string
child 4, query_unchanged: bool
child 5, inputs_verified: bool
child 6, source_versions_verified: bool
child 7, raw_evidence_hashes: struct<commands.jsonl: string, snapshots.jsonl: string, renders.jsonl: string, error.jsonl: string, (... 74 chars omitted)
child 0, commands.jsonl: string
child 1, snapshots.jsonl: string
child 2, renders.jsonl: string
child 3, error.jsonl: string
child 4, image_views.jsonl: string
child 5, artifacts.jsonl: string
child 6, rollout.jsonl: string
child 8, audit: struct<task_id: string, collection_integrity_ok: bool, browser_ok: bool, roundtrip_ok: bool, issues: (... 6509 chars omitted)
child 0, task_id: string
child 1, collection_integrity_ok: bool
child 2, browser_ok: bool
child 3, roundtrip_ok: bool
child 4, issues: list<item: null>
child 0, item: null
child 5
...
_03: string
child 207, transport_turn_0001_1: string
child 208, transport_turn_0001_2: string
child 209, trail_surface_alignment: string
child 210, export_unused_path_vertices: string
child 211, terrain_visual: string
child 212, forest_composition: string
child 213, river_visibility: string
child 214, roundtrip_cmd_000029: string
child 215, delivery_cmd_000029: string
child 216, browser_volume: string
child 217, roundtrip_cmd_000040: string
child 218, delivery_cmd_000040: string
child 219, browser_json_limit: string
child 220, explicit_resume_2026-09-24T02:09:44Z: string
child 221, composition_occlusion: string
child 222, water_surface_overlap: string
child 223, master_frame_margin: string
child 224, backdrop_edge: string
child 225, door_occlusion: string
child 226, bench_occlusion: string
child 227, HOLO-01: string
child 228, FRAME-02: string
child 229, browser_facade_artifacts: string
child 21, semantic_quality_status: string
child 22, formal_benchmark_pass: null
child 23, ready_for_review: bool
source_files_verified: int64
scope: string
selected_ids: list<item: string>
child 0, item: string
credential_pattern_scan_hits: int64
to
{'scope': Value('string'), 'selected_ids': List(Value('string')), 'source_revision': Value('string'), 'source_files_verified': Value('int64'), 'credential_pattern_scan_hits': Value('int64'), 'results': List({'id': Value('string'), 'run': Value('string'), 'audit': {'task_id': Value('string'), 'collection_integrity_ok': Value('bool'), 'browser_ok': Value('bool'), 'roundtrip_ok': Value('bool'), 'issues': List(Value('null')), 'command_count': Value('int64'), 'command_snapshots': Value('int64'), 'render_count': Value('int64'), 'image_view_count': Value('int64'), 'model_image_views': Value('int64'), 'model_render_views': Value('int64'), 'model_visual_review_complete': Value('bool'), 'native_image_outputs': Value('int64'), 'native_unique_model_images': Value('int64'), 'model_unique_requested_images': Value('int64'), 'current_turn': Value('string'), 'current_turn_visual_review_complete': Value('bool'), 'current_turn_model_commands': Value('int64'), 'error_recording': Value('string'), 'error_count': Value('int64'), 'issue_latest_states': {'error_cmd_000003': Value('string'), 'error_cmd_000004': Value('string'), 'composition_pool_01': Value('string'), 'composition_01': Value('string'), 'error_cmd_000008': Value('string'), 'backdrop_01': Value('string'), 'error_cmd_000010': Value('string'), 'bridge_approaches': Value('string'), 'roof_surface': Value('string'), 'browser_surface_artifacts': Value('string')}, 'semantic_quality_status': Value('string'), 'formal_benchmark_pass': Value('null'), 'ready_for_review': Value('bool')}, 'artifacts': {'final.blend': {'bytes': Value('int64'), 'sha256': Value('string')}, 'final.glb': {'bytes': Value('int64'), 'sha256': Value('string')}, 'scene.json': {'bytes': Value('int64'), 'sha256': Value('string')}, 'viewer.html': {'bytes': Value('int64'), 'sha256': Value('string')}, 'browser.png': {'bytes': Value('int64'), 'sha256': Value('string')}}, 'query_unchanged': Value('bool'), 'input_hashes_verified': 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 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
count: int64
unique_ids: int64
unique_sessions: int64
all_engineering_passed: bool
source_revision: string
semantic_quality: string
results: list<item: struct<id: string, batch: string, run: string, thread_id: string, query_unchanged: bool, (... 6870 chars omitted)
child 0, item: struct<id: string, batch: string, run: string, thread_id: string, query_unchanged: bool, inputs_veri (... 6858 chars omitted)
child 0, id: string
child 1, batch: string
child 2, run: string
child 3, thread_id: string
child 4, query_unchanged: bool
child 5, inputs_verified: bool
child 6, source_versions_verified: bool
child 7, raw_evidence_hashes: struct<commands.jsonl: string, snapshots.jsonl: string, renders.jsonl: string, error.jsonl: string, (... 74 chars omitted)
child 0, commands.jsonl: string
child 1, snapshots.jsonl: string
child 2, renders.jsonl: string
child 3, error.jsonl: string
child 4, image_views.jsonl: string
child 5, artifacts.jsonl: string
child 6, rollout.jsonl: string
child 8, audit: struct<task_id: string, collection_integrity_ok: bool, browser_ok: bool, roundtrip_ok: bool, issues: (... 6509 chars omitted)
child 0, task_id: string
child 1, collection_integrity_ok: bool
child 2, browser_ok: bool
child 3, roundtrip_ok: bool
child 4, issues: list<item: null>
child 0, item: null
child 5
...
_03: string
child 207, transport_turn_0001_1: string
child 208, transport_turn_0001_2: string
child 209, trail_surface_alignment: string
child 210, export_unused_path_vertices: string
child 211, terrain_visual: string
child 212, forest_composition: string
child 213, river_visibility: string
child 214, roundtrip_cmd_000029: string
child 215, delivery_cmd_000029: string
child 216, browser_volume: string
child 217, roundtrip_cmd_000040: string
child 218, delivery_cmd_000040: string
child 219, browser_json_limit: string
child 220, explicit_resume_2026-09-24T02:09:44Z: string
child 221, composition_occlusion: string
child 222, water_surface_overlap: string
child 223, master_frame_margin: string
child 224, backdrop_edge: string
child 225, door_occlusion: string
child 226, bench_occlusion: string
child 227, HOLO-01: string
child 228, FRAME-02: string
child 229, browser_facade_artifacts: string
child 21, semantic_quality_status: string
child 22, formal_benchmark_pass: null
child 23, ready_for_review: bool
source_files_verified: int64
scope: string
selected_ids: list<item: string>
child 0, item: string
credential_pattern_scan_hits: int64
to
{'scope': Value('string'), 'selected_ids': List(Value('string')), 'source_revision': Value('string'), 'source_files_verified': Value('int64'), 'credential_pattern_scan_hits': Value('int64'), 'results': List({'id': Value('string'), 'run': Value('string'), 'audit': {'task_id': Value('string'), 'collection_integrity_ok': Value('bool'), 'browser_ok': Value('bool'), 'roundtrip_ok': Value('bool'), 'issues': List(Value('null')), 'command_count': Value('int64'), 'command_snapshots': Value('int64'), 'render_count': Value('int64'), 'image_view_count': Value('int64'), 'model_image_views': Value('int64'), 'model_render_views': Value('int64'), 'model_visual_review_complete': Value('bool'), 'native_image_outputs': Value('int64'), 'native_unique_model_images': Value('int64'), 'model_unique_requested_images': Value('int64'), 'current_turn': Value('string'), 'current_turn_visual_review_complete': Value('bool'), 'current_turn_model_commands': Value('int64'), 'error_recording': Value('string'), 'error_count': Value('int64'), 'issue_latest_states': {'error_cmd_000003': Value('string'), 'error_cmd_000004': Value('string'), 'composition_pool_01': Value('string'), 'composition_01': Value('string'), 'error_cmd_000008': Value('string'), 'backdrop_01': Value('string'), 'error_cmd_000010': Value('string'), 'bridge_approaches': Value('string'), 'roof_surface': Value('string'), 'browser_surface_artifacts': Value('string')}, 'semantic_quality_status': Value('string'), 'formal_benchmark_pass': Value('null'), 'ready_for_review': Value('bool')}, 'artifacts': {'final.blend': {'bytes': Value('int64'), 'sha256': Value('string')}, 'final.glb': {'bytes': Value('int64'), 'sha256': Value('string')}, 'scene.json': {'bytes': Value('int64'), 'sha256': Value('string')}, 'viewer.html': {'bytes': Value('int64'), 'sha256': Value('string')}, 'browser.png': {'bytes': Value('int64'), 'sha256': Value('string')}}, 'query_unchanged': Value('bool'), 'input_hashes_verified': 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.
Astra3D split_01 pilot trajectories
Collector: NathanTong / Zhao TONG. Three pilot queries sampled on 2026-09-23 (Asia/Shanghai), using gpt-6-astra with medium reasoning and Blender 4.2.23 on Windows.
Source queries: yasNing/VibeWorlding-V2-Query, split_01. Pipeline: yasNing/Astra3D-Bench sampling, with Windows changes and sampling-time source snapshot included in the archive.
| ID | Input coverage | End-to-end time | Commands/snapshots |
|---|---|---|---|
| Q0095 | Text | 4m56s | 19/19 |
| Q0427 | Text and reference image | 8m30s | 28/28 |
| Q0058 | Text and supplied GLB assets | 13m28s | 50/50 |
All three passed collection integrity, browser and structural GLB roundtrip checks. Semantic quality remains pending review; engineering completion is not benchmark acceptance. Known visual limitations and independent observations are included. Timing includes model/tool/network/rendering waits and retries, not pure model inference. The remaining 97 source queries have not been sampled.
The archive preserves original trajectories, all command snapshots, intermediate renders, errors and repairs, original and transported images, final Blender/GLB files, offline viewer dependencies, source/configuration provenance, and supplementary inspection checks. Files are losslessly deduplicated by identical bytes; events or scene iterations are never omitted. Full paths in historical logs are retained.
Restore and view
This is a content-addressed evidence archive, not a regular ready-to-browse ZIP. Download the archive, restore.py and evidence_io.py into one directory. Python 3.9+ is sufficient to restore; Blender is only needed to edit native scenes.
python restore.py verify pilot01-full-evidence.zip
python restore.py restore pilot01-full-evidence.zip restored
cd restored/pilot01
python -m http.server 8000 --bind 127.0.0.1
Open http://127.0.0.1:8000/review/inspection/index.html for the comparison page, or http://127.0.0.1:8000/review/index.html for collection records. Use python3 instead of python if that is your installed command. Restore destination must not already exist. Do not open the objects directory as a viewer.
See PILOT_REPORT.md, timing_summary.json and delivery_verification.json for details. Independent restored-viewer checks are supplementary verification, not part of model behavior. Original quality_review.json records are not changed into passing scores.
This repository is intended as a private research handoff. It makes no new license grant for upstream query, image, asset or pipeline material.
- Downloads last month
- 151