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
status: string
scope: string
schema_v1_admitted: bool
all_storage_v3: bool
original_episode_count: int64
reviewed_clip_count: int64
episodes_by_task: struct<nut_thread: int64, pc_gpu: int64, pen_holder: int64, tool_packing: int64, coffee: int64, pc_r (... 29 chars omitted)
child 0, nut_thread: int64
child 1, pc_gpu: int64
child 2, pen_holder: int64
child 3, tool_packing: int64
child 4, coffee: int64
child 5, pc_ram: int64
child 6, pc_gpu_ram: int64
dataset_count: int64
episode_index_rows: int64
primitive_index_rows: int64
blacklist_rows: int64
bidirectional_links_verified: int64
video_reencoding: bool
recrop: bool
derived_conversion: list<item: struct<task: string, source_codebase_version: string, codebase_version: string, episode_c (... 382 chars omitted)
child 0, item: struct<task: string, source_codebase_version: string, codebase_version: string, episode_count: int64 (... 370 chars omitted)
child 0, task: string
child 1, source_codebase_version: string
child 2, codebase_version: string
child 3, episode_count: int64
child 4, frame_count: int64
child 5, video_count: int64
child 6, all_numeric_values_read_back: bool
child 7, all_numeric_values_unchanged: bool
child 8, renumbered_columns: list<item: null>
child 0, item: null
child 9, video_paths_and_bytes_preserved: bool
child 10, source_video_sha256_recorded: bool
child 11, video_decoding_performed: bool
child 12, video_enc
...
item: string
child 5, schema_v1_admitted: bool
child 6, data_license: string
bidirectional_annotations_checked: int64
missing_source_fields: list<item: string>
child 0, item: string
leaves: list<item: struct<task: string, path: string, episodes: int64, clips: int64, native_validation: stru (... 430 chars omitted)
child 0, item: struct<task: string, path: string, episodes: int64, clips: int64, native_validation: struct<complete (... 418 chars omitted)
child 0, task: string
child 1, path: string
child 2, episodes: int64
child 3, clips: int64
child 4, native_validation: struct<complete_source_episodes: bool, all_numeric_values_read_back: bool, native_values_preserved_e (... 332 chars omitted)
child 0, complete_source_episodes: bool
child 1, all_numeric_values_read_back: bool
child 2, native_values_preserved_except_format_indices: bool
child 3, renumbered_columns: list<item: string>
child 0, item: string
child 4, preserved_columns: list<item: string>
child 0, item: string
child 5, source_file_sha256_verified: bool
child 6, video_storage: string
child 7, video_decoding_performed: bool
child 8, video_statistics: string
child 9, global_statistics: string
child 10, official_reader: string
child 11, full_schema_compliance: bool
child 12, global_index_created: bool
formal_index_generated: bool
to
{'status': Value('string'), 'scope': Value('string'), 'schema_v1_admitted': Value('bool'), 'original_episode_count': Value('int64'), 'reviewed_clip_count': Value('int64'), 'episodes_by_task': {'nut_thread': Value('int64'), 'pc_gpu': Value('int64'), 'pen_holder': Value('int64'), 'tool_packing': Value('int64')}, 'leaves': List({'task': Value('string'), 'path': Value('string'), 'episodes': Value('int64'), 'clips': Value('int64'), 'native_validation': {'complete_source_episodes': Value('bool'), 'all_numeric_values_read_back': Value('bool'), 'native_values_preserved_except_format_indices': Value('bool'), 'renumbered_columns': List(Value('string')), 'preserved_columns': List(Value('string')), 'source_file_sha256_verified': Value('bool'), 'video_storage': Value('string'), 'video_decoding_performed': Value('bool'), 'video_statistics': Value('string'), 'global_statistics': Value('string'), 'official_reader': Value('string'), 'full_schema_compliance': Value('bool'), 'global_index_created': Value('bool')}}), 'bidirectional_annotations_checked': Value('int64'), 'formal_index_generated': Value('bool'), 'missing_source_fields': List(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 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
status: string
scope: string
schema_v1_admitted: bool
all_storage_v3: bool
original_episode_count: int64
reviewed_clip_count: int64
episodes_by_task: struct<nut_thread: int64, pc_gpu: int64, pen_holder: int64, tool_packing: int64, coffee: int64, pc_r (... 29 chars omitted)
child 0, nut_thread: int64
child 1, pc_gpu: int64
child 2, pen_holder: int64
child 3, tool_packing: int64
child 4, coffee: int64
child 5, pc_ram: int64
child 6, pc_gpu_ram: int64
dataset_count: int64
episode_index_rows: int64
primitive_index_rows: int64
blacklist_rows: int64
bidirectional_links_verified: int64
video_reencoding: bool
recrop: bool
derived_conversion: list<item: struct<task: string, source_codebase_version: string, codebase_version: string, episode_c (... 382 chars omitted)
child 0, item: struct<task: string, source_codebase_version: string, codebase_version: string, episode_count: int64 (... 370 chars omitted)
child 0, task: string
child 1, source_codebase_version: string
child 2, codebase_version: string
child 3, episode_count: int64
child 4, frame_count: int64
child 5, video_count: int64
child 6, all_numeric_values_read_back: bool
child 7, all_numeric_values_unchanged: bool
child 8, renumbered_columns: list<item: null>
child 0, item: null
child 9, video_paths_and_bytes_preserved: bool
child 10, source_video_sha256_recorded: bool
child 11, video_decoding_performed: bool
child 12, video_enc
...
item: string
child 5, schema_v1_admitted: bool
child 6, data_license: string
bidirectional_annotations_checked: int64
missing_source_fields: list<item: string>
child 0, item: string
leaves: list<item: struct<task: string, path: string, episodes: int64, clips: int64, native_validation: stru (... 430 chars omitted)
child 0, item: struct<task: string, path: string, episodes: int64, clips: int64, native_validation: struct<complete (... 418 chars omitted)
child 0, task: string
child 1, path: string
child 2, episodes: int64
child 3, clips: int64
child 4, native_validation: struct<complete_source_episodes: bool, all_numeric_values_read_back: bool, native_values_preserved_e (... 332 chars omitted)
child 0, complete_source_episodes: bool
child 1, all_numeric_values_read_back: bool
child 2, native_values_preserved_except_format_indices: bool
child 3, renumbered_columns: list<item: string>
child 0, item: string
child 4, preserved_columns: list<item: string>
child 0, item: string
child 5, source_file_sha256_verified: bool
child 6, video_storage: string
child 7, video_decoding_performed: bool
child 8, video_statistics: string
child 9, global_statistics: string
child 10, official_reader: string
child 11, full_schema_compliance: bool
child 12, global_index_created: bool
formal_index_generated: bool
to
{'status': Value('string'), 'scope': Value('string'), 'schema_v1_admitted': Value('bool'), 'original_episode_count': Value('int64'), 'reviewed_clip_count': Value('int64'), 'episodes_by_task': {'nut_thread': Value('int64'), 'pc_gpu': Value('int64'), 'pen_holder': Value('int64'), 'tool_packing': Value('int64')}, 'leaves': List({'task': Value('string'), 'path': Value('string'), 'episodes': Value('int64'), 'clips': Value('int64'), 'native_validation': {'complete_source_episodes': Value('bool'), 'all_numeric_values_read_back': Value('bool'), 'native_values_preserved_except_format_indices': Value('bool'), 'renumbered_columns': List(Value('string')), 'preserved_columns': List(Value('string')), 'source_file_sha256_verified': Value('bool'), 'video_storage': Value('string'), 'video_decoding_performed': Value('bool'), 'video_statistics': Value('string'), 'global_statistics': Value('string'), 'official_reader': Value('string'), 'full_schema_compliance': Value('bool'), 'global_index_created': Value('bool')}}), 'bidirectional_annotations_checked': Value('int64'), 'formal_index_generated': Value('bool'), 'missing_source_fields': List(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.
EmbodiedSWE：统一 LeRobot v3 与 Schema 目录
当前版本在 **organized-v2/data/**：
data/
registry/ primitives.yaml tasks.yaml embodiments.yaml experts.yaml splits.yaml
quality/ blacklist.parquet
index/ episodes.parquet primitives.parquet
sim/{source}/{embodiment_id}/{task_family}/
meta/ data/ videos/ segments.parquet
| source | 内容 | 数量 |
|---|---|---|
| embodiedswe | 完整原始 episode | 133 |
| embodiedswe_cropped | 裁后 episode，四类各50段 | 200 |
| isaac_a2r_swe | 仿真专家完整 episode | 150 |
| isaac_a2r_swe_cropped | 仿真专家裁后片段 | 150 |
共14个叶子数据集，均使用 LeRobot v3。任务和本体匹配包内注册表；未注册任务须用 _unregistered。目录维度与元数据一致。
索引由脚本生成：633个episode、700条phase（原始与裁后分别记录，不代表等量的独立交互）。用 dataset_role 和 parent_episode_id 区分与关联。原始和裁后 JSON 双向关联；视频、数值、时间戳及已确认裁剪边界保持不变。
源缺失字段如实标记并列入质量表，未补造原生测量。目录与存储格式统一不等于训练准入，索引中的 schema_v1_admitted 保持实际状态。
验收结果 · 版本与转换记录。organized-v1/、release/ 与 runs/ 保留作历史记录，当前请查看 organized-v2/。
仿真专家数据（source=isaac_a2r_swe）
由本项目 EmbodiedSWE 接触专家在 Isaac Sim 中生成，本体 franka_swe_diff_ik（Franka，末端增量动作，见注册表）。每个任务取通过物理验收且裁剪经复核的前 50 个 seed；完整轨迹在 isaac_a2r_swe，裁后片段在 isaac_a2r_swe_cropped。
| 任务 | Hz | 完整 episode | 裁后片段 | 帧数（完整 / 裁后） | 逐条复核 |
|---|---|---|---|---|---|
| coffee | 60 | 50 | 50 | 24710 / 17200 | 13 |
| pc_ram | 48 | 50 | 50 | 10850 / 10300 | 12 |
| pc_gpu_ram | 48 | 50 | 50 | 10851 / 10297 | 12 |
- 与上游开源数据不同，含
observation.ee_pose、control_source、sim_seed、episode_randomization和privileged.*仿真真值；privileged.*只用于标注、判定与分析，不作为策略输入。 - 完整轨迹的数值与视频字节沿用原生转换结果，只重编
episode_index/index。裁后片段的数值取自父 episode 对应行，frame_index/timestamp从 0 重排；裁后视频在裁剪导出时由父视频重新编码为 H.264，本次合并未再编码。 - 裁剪边界：每任务先复核一个模板，按 DTW 或实测事件传播到其余 episode，再分层抽查（“逐条复核”列为抽查条数），
reviews/逐条写明依据。片段级物理成功与critical未标注。 coffee：只做空手推动已打开的咖啡机滑盖使其关闭，不含冲煮或端杯。pc_ram：从已持内存条的局部起点插入 0 号槽、松开并撤离，不含拿取。pc_gpu_ram：显卡在场时插入 0 号内存条、松开并撤离，不插显卡，不代表整机装配成功。- 数据许可
apache-2.0，由数据所有者在本次发布时指定；采集时记录的内部许可保留在collection_license。仿真场景与物体素材沿用上游各自的许可。 - 这些 episode 同样
schema_v1_admitted=false。
- Downloads last month
- 451