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
task: string
embodiment: string
episode: int64
seed: int64
frames: int64
cameras: list<item: string>
child 0, item: string
hdf5: string
hdf5_bytes: int64
replay_verified: bool
task_success: bool
video_frames_verified: bool
arrays_verified: bool
canonical_rest_id: string
canonical_policy: string
source_hdf5_sha256: string
max_link_local_error_m: double
view_frames: int64
checks: list<item: string>
child 0, item: string
counts: struct<lift_pot/ARX-X5: int64, lift_pot/aloha-agilex: int64, lift_pot/franka-panda: int64, lift_pot/ (... 205 chars omitted)
child 0, lift_pot/ARX-X5: int64
child 1, lift_pot/aloha-agilex: int64
child 2, lift_pot/franka-panda: int64
child 3, lift_pot/piper: int64
child 4, lift_pot/ur5-wsg: int64
child 5, place_empty_cup/ARX-X5: int64
child 6, place_empty_cup/aloha-agilex: int64
child 7, place_empty_cup/franka-panda: int64
child 8, place_empty_cup/piper: int64
child 9, place_empty_cup/ur5-wsg: int64
episodes: int64
videos: int64
rest_ids: struct<ARX-X5: string, aloha-agilex: string, franka-panda: string, piper: string, ur5-wsg: string>
child 0, ARX-X5: string
child 1, aloha-agilex: string
child 2, franka-panda: string
child 3, piper: string
child 4, ur5-wsg: string
camera_sequences: int64
all_checks_passed: bool
to
{'episodes': Value('int64'), 'videos': Value('int64'), 'frames': Value('int64'), 'camera_sequences': Value('int64'), 'view_frames': Value('int64'), 'hdf5_bytes': Value('int64'), 'max_link_local_error_m': Value('float64'), 'counts': {'lift_pot/ARX-X5': Value('int64'), 'lift_pot/aloha-agilex': Value('int64'), 'lift_pot/franka-panda': Value('int64'), 'lift_pot/piper': Value('int64'), 'lift_pot/ur5-wsg': Value('int64'), 'place_empty_cup/ARX-X5': Value('int64'), 'place_empty_cup/aloha-agilex': Value('int64'), 'place_empty_cup/franka-panda': Value('int64'), 'place_empty_cup/piper': Value('int64'), 'place_empty_cup/ur5-wsg': Value('int64')}, 'rest_ids': {'ARX-X5': Value('string'), 'aloha-agilex': Value('string'), 'franka-panda': Value('string'), 'piper': Value('string'), 'ur5-wsg': Value('string')}, 'all_checks_passed': Value('bool'), 'checks': 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 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
task: string
embodiment: string
episode: int64
seed: int64
frames: int64
cameras: list<item: string>
child 0, item: string
hdf5: string
hdf5_bytes: int64
replay_verified: bool
task_success: bool
video_frames_verified: bool
arrays_verified: bool
canonical_rest_id: string
canonical_policy: string
source_hdf5_sha256: string
max_link_local_error_m: double
view_frames: int64
checks: list<item: string>
child 0, item: string
counts: struct<lift_pot/ARX-X5: int64, lift_pot/aloha-agilex: int64, lift_pot/franka-panda: int64, lift_pot/ (... 205 chars omitted)
child 0, lift_pot/ARX-X5: int64
child 1, lift_pot/aloha-agilex: int64
child 2, lift_pot/franka-panda: int64
child 3, lift_pot/piper: int64
child 4, lift_pot/ur5-wsg: int64
child 5, place_empty_cup/ARX-X5: int64
child 6, place_empty_cup/aloha-agilex: int64
child 7, place_empty_cup/franka-panda: int64
child 8, place_empty_cup/piper: int64
child 9, place_empty_cup/ur5-wsg: int64
episodes: int64
videos: int64
rest_ids: struct<ARX-X5: string, aloha-agilex: string, franka-panda: string, piper: string, ur5-wsg: string>
child 0, ARX-X5: string
child 1, aloha-agilex: string
child 2, franka-panda: string
child 3, piper: string
child 4, ur5-wsg: string
camera_sequences: int64
all_checks_passed: bool
to
{'episodes': Value('int64'), 'videos': Value('int64'), 'frames': Value('int64'), 'camera_sequences': Value('int64'), 'view_frames': Value('int64'), 'hdf5_bytes': Value('int64'), 'max_link_local_error_m': Value('float64'), 'counts': {'lift_pot/ARX-X5': Value('int64'), 'lift_pot/aloha-agilex': Value('int64'), 'lift_pot/franka-panda': Value('int64'), 'lift_pot/piper': Value('int64'), 'lift_pot/ur5-wsg': Value('int64'), 'place_empty_cup/ARX-X5': Value('int64'), 'place_empty_cup/aloha-agilex': Value('int64'), 'place_empty_cup/franka-panda': Value('int64'), 'place_empty_cup/piper': Value('int64'), 'place_empty_cup/ur5-wsg': Value('int64')}, 'rest_ids': {'ARX-X5': Value('string'), 'aloha-agilex': Value('string'), 'franka-panda': Value('string'), 'piper': Value('string'), 'ur5-wsg': Value('string')}, 'all_checks_passed': Value('bool'), 'checks': 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.
RoboTwin 2.0 multi-view CCM — fixed rest
RoboTwin 官方轨迹经仿真 replay 生成的多视角 RGB / CCM 数据，每种机器人共用固定 canonical 参考姿态。
数据规模
- 两个任务：
place_empty_cup、lift_pot。 - 五种本体：Aloha-AgileX、ARX-X5、Franka Panda、Piper、UR5-WSG。
- 每个任务／本体各 10 条，共 100 条 episode、10,922 个时间帧。
- 320 路相机序列、35,622 个视角帧；Aloha 四视角，其余本体三视角。
- 100 个 HDF5、各视角浮点 NPZ、canonical GLB/JSON、2,240 个预览和诊断视频。
- 压缩包 5,909,867,502 bytes（约 5.50 GiB）；解压后文件总大小约 7.13 GiB。 保留压缩包时建议至少预留 15 GiB 空间。
下载与解压
下载本仓库的压缩包及 SHA256SUMS 后运行：
sha256sum -c SHA256SUMS
mkdir -p extracted
tar -xzf robotwin2_two_tasks_fixed_rest.tar.gz -C extracted
解压根目录为 extracted/robotwin2_two_tasks_fixed_rest/。
dataset_index_portable.json 中 hdf5 为相对于该根目录的路径。
包内原始索引和部分元数据保留生成机器的绝对路径，仅作来源记录；在新机器上请使用相对路径索引。
import json
from pathlib import Path
import h5py
root = Path("extracted/robotwin2_two_tasks_fixed_rest")
records = json.loads(Path("dataset_index_portable.json").read_text())
with h5py.File(root / records[0]["hdf5"], "r") as f:
print(list(f.keys()))
print(list(f["observation"].keys()))
坐标与控制语义
每个本体以 place_empty_cup/episode0 首帧实际姿态建立固定 rest。
全部任务、episode 和相机使用同一份本体参考定义；每路相机均附 GLB 和 canonical sidecar 副本，
使用这些副本无需另行下载生成机器上的共享 registry。
每路 HDF5 包含 RGB、浮点 CCM、有效 mask、link ID、米制 depth 和相机标定。 CCM 无效像素为 NaN，link ID 背景为 0；一些腕部视角可能全程看不到机器人，应按有效 mask 处理。 MP4 用于可视化，训练使用 HDF5/NPZ 浮点标签。
robot_state/qpos_actual 保存物理关节位置，canonical/reference_qpos 保存固定 rest。
joint_action 保留采样时刻的官方控制语义，不能将其等同于 actual qpos 或相邻帧关节差分。
物理时间由记录的 physics tick 和 timestep 确定，不应从视频 FPS 推导。
该固定 rest 版本对已验收的 CCM 做逐 link 参考坐标变换，保持 RGB、mask、depth、
相机、actual qpos、控制和时间不变，没有重新运行物理仿真。
最大 link-local 坐标差为 2.3904972556998416e-08 m，表示坐标变换舍入误差，
不表示机器人标注或 action 恢复误差。验收详情见包内及仓库附带的 audit_report.json。
代码、文档及来源
来源数据集：TianxingChen/RoboTwin2.0。 本数据是派生 CCM 标注，非 RoboTwin 官方发布。上游数据与机器人资产的使用条件仍需遵守， 此说明不为上游内容重新授权。这里发布的是观测／标注数据，不包含完整仿真资产或全部场景恢复文件。
- Downloads last month
- 78