Dataset Viewer

Split (1)

train

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
collection: string
created_date: timestamp[s]
source_roots: list<item: string>
  child 0, item: string
selection: struct<rule: string, candidate_count: int64, episode_count: int64, count_by_task: struct<fold_towels (... 177 chars omitted)
  child 0, rule: string
  child 1, candidate_count: int64
  child 2, episode_count: int64
  child 3, count_by_task: struct<fold_towels1: int64, fold_towels: int64, fold_towels2: int64>
      child 0, fold_towels1: int64
      child 1, fold_towels: int64
      child 2, fold_towels2: int64
  child 4, total_duration_seconds: double
  child 5, shortest_selected_seconds: double
  child 6, longest_selected_seconds: double
  child 7, total_mcap_bytes: int64
raw_data_policy: struct<mcap_unchanged: bool, image_reencoding: bool, resize: bool, downsampling: bool, normalization (... 7 chars omitted)
  child 0, mcap_unchanged: bool
  child 1, image_reencoding: bool
  child 2, resize: bool
  child 3, downsampling: bool
  child 4, normalization: bool
episodes: list<item: struct<rank_by_duration: int64, task_directory: string, task: string, episode: string, du (... 915 chars omitted)
  child 0, item: struct<rank_by_duration: int64, task_directory: string, task: string, episode: string, duration_seco (... 903 chars omitted)
      child 0, rank_by_duration: int64
      child 1, task_directory: string
      child 2, task: string
      child 3, episode: string
      child 4, duration_seconds: double
      child 5, success: null
      child 6, normal_source_dire
...
  child 13, /master/joint_right: int64
          child 14, /puppet/end_pose_left: int64
          child 15, /puppet/end_pose_right: int64
          child 16, /puppet/joint_left: int64
          child 17, /puppet/joint_right: int64
          child 18, /tf: int64
          child 19, /tf_static: int64
      child 13, chunk_crc_validated: bool
      child 14, required_topics_present: bool
data_info: struct<library: string, profile: string, messages: int64, start_time: string, start_timestamp: doubl (... 217 chars omitted)
  child 0, library: string
  child 1, profile: string
  child 2, messages: int64
  child 3, start_time: string
  child 4, start_timestamp: double
  child 5, end_time: string
  child 6, end_timestamp: double
  child 7, channels: int64
  child 8, attachments: int64
  child 9, metadata: int64
  child 10, topics: list<item: struct<topic: string, count: int64, frequency: double, type: string, serialization_format (... 10 chars omitted)
      child 0, item: struct<topic: string, count: int64, frequency: double, type: string, serialization_format: string>
          child 0, topic: string
          child 1, count: int64
          child 2, frequency: double
          child 3, type: string
          child 4, serialization_format: string
task_info: struct<episode: int64, task_id: string, task_name: string, file_name: string, steps: string>
  child 0, episode: int64
  child 1, task_id: string
  child 2, task_name: string
  child 3, file_name: string
  child 4, steps: string
to
{'data_info': {'library': Value('string'), 'profile': Value('string'), 'messages': Value('int64'), 'start_time': Value('string'), 'start_timestamp': Value('float64'), 'end_time': Value('string'), 'end_timestamp': Value('float64'), 'channels': Value('int64'), 'attachments': Value('int64'), 'metadata': Value('int64'), 'topics': List({'topic': Value('string'), 'count': Value('int64'), 'frequency': Value('float64'), 'type': Value('string'), 'serialization_format': Value('string')})}, 'task_info': {'episode': Value('int64'), 'task_id': Value('string'), 'task_name': Value('string'), 'file_name': Value('string'), 'steps': Value('string')}}
because column names don't match
Traceback:    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 489, 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 2818, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2355, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2380, 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 2369, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2297, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              collection: string
              created_date: timestamp[s]
              source_roots: list<item: string>
                child 0, item: string
              selection: struct<rule: string, candidate_count: int64, episode_count: int64, count_by_task: struct<fold_towels (... 177 chars omitted)
                child 0, rule: string
                child 1, candidate_count: int64
                child 2, episode_count: int64
                child 3, count_by_task: struct<fold_towels1: int64, fold_towels: int64, fold_towels2: int64>
                    child 0, fold_towels1: int64
                    child 1, fold_towels: int64
                    child 2, fold_towels2: int64
                child 4, total_duration_seconds: double
                child 5, shortest_selected_seconds: double
                child 6, longest_selected_seconds: double
                child 7, total_mcap_bytes: int64
              raw_data_policy: struct<mcap_unchanged: bool, image_reencoding: bool, resize: bool, downsampling: bool, normalization (... 7 chars omitted)
                child 0, mcap_unchanged: bool
                child 1, image_reencoding: bool
                child 2, resize: bool
                child 3, downsampling: bool
                child 4, normalization: bool
              episodes: list<item: struct<rank_by_duration: int64, task_directory: string, task: string, episode: string, du (... 915 chars omitted)
                child 0, item: struct<rank_by_duration: int64, task_directory: string, task: string, episode: string, duration_seco (... 903 chars omitted)
                    child 0, rank_by_duration: int64
                    child 1, task_directory: string
                    child 2, task: string
                    child 3, episode: string
                    child 4, duration_seconds: double
                    child 5, success: null
                    child 6, normal_source_dire
              ...
                child 13, /master/joint_right: int64
                        child 14, /puppet/end_pose_left: int64
                        child 15, /puppet/end_pose_right: int64
                        child 16, /puppet/joint_left: int64
                        child 17, /puppet/joint_right: int64
                        child 18, /tf: int64
                        child 19, /tf_static: int64
                    child 13, chunk_crc_validated: bool
                    child 14, required_topics_present: bool
              data_info: struct<library: string, profile: string, messages: int64, start_time: string, start_timestamp: doubl (... 217 chars omitted)
                child 0, library: string
                child 1, profile: string
                child 2, messages: int64
                child 3, start_time: string
                child 4, start_timestamp: double
                child 5, end_time: string
                child 6, end_timestamp: double
                child 7, channels: int64
                child 8, attachments: int64
                child 9, metadata: int64
                child 10, topics: list<item: struct<topic: string, count: int64, frequency: double, type: string, serialization_format (... 10 chars omitted)
                    child 0, item: struct<topic: string, count: int64, frequency: double, type: string, serialization_format: string>
                        child 0, topic: string
                        child 1, count: int64
                        child 2, frequency: double
                        child 3, type: string
                        child 4, serialization_format: string
              task_info: struct<episode: int64, task_id: string, task_name: string, file_name: string, steps: string>
                child 0, episode: int64
                child 1, task_id: string
                child 2, task_name: string
                child 3, file_name: string
                child 4, steps: string
              to
              {'data_info': {'library': Value('string'), 'profile': Value('string'), 'messages': Value('int64'), 'start_time': Value('string'), 'start_timestamp': Value('float64'), 'end_time': Value('string'), 'end_timestamp': Value('float64'), 'channels': Value('int64'), 'attachments': Value('int64'), 'metadata': Value('int64'), 'topics': List({'topic': Value('string'), 'count': Value('int64'), 'frequency': Value('float64'), 'type': Value('string'), 'serialization_format': Value('string')})}, 'task_info': {'episode': Value('int64'), 'task_id': Value('string'), 'task_name': Value('string'), 'file_name': Value('string'), 'steps': Value('string')}}
              because column names don't match

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

真机原始轨迹样本（2026-07-20）

本目录从远程 OSS oss://ss-oss1/data/dataset/external_robotic_data/ 收集了 4 条正常目录下的原始真机轨迹。数据保持原始 MCAP/JPEG/ROS2 物理量和时间戳，未做归一化、拼图、resize、降采样或视频重编码。

样本

任务	Episode	时长	MCAP 大小	RGB	控制/状态
制作咖啡	1	73.776 s	1,568,867,100 B	4 路，1280×720，约 30 FPS	双臂约 200 Hz
叠紫色毛巾-重制版	1	43.094 s	957,553,080 B	4 路，1280×720，约 30 FPS	双臂约 200 Hz
制作三明治	1	43.478 s	945,035,379 B	4 路，1280×720，约 30 FPS	双臂约 200 Hz
开关台灯	1	27.477 s	420,049,075 B	4 路，1280×720，约 30 FPS	双臂约 200 Hz

原始 episode 位于 episodes/，任务级 instruction/annotation 位于 task_metadata/。逐文件 SHA-256 见 SHA256SUMS，机器可读清单见 manifest.json。

文档字段映射

RGB

head_camera / 顶部正面：/camera_f/color/image_raw
left_wrist_camera：/camera_l/color/image_raw
right_wrist_camera：/camera_r/color/image_raw
额外正面左侧相机：/camera_fl/color/image_raw
图像编码：原始 JPEG sensor_msgs/msg/CompressedImage
原始分辨率：四路均为 1280×720
时间戳：每条消息同时保留 ROS header timestamp 与 MCAP log/publish timestamp，单位为纳秒
相机内参：对应 /camera_{f,l,r,fl}/color/camera_info
相机外参：/tf_static 明确包含 camera_front、camera_front_left、camera_left_wrist、camera_right_wrist

Action

最佳原始对应为 /master/joint_left 和 /master/joint_right：控制器收到的目标关节“指令回显”，约 200 Hz。

类型：absolute joint position
每臂原始消息：[joint0, joint1, joint2, joint3, joint4, joint5, gripper]
统一 14D 顺序：[L_j0..L_j5, R_j0..R_j5, L_gripper, R_gripper]
六个臂关节单位：rad
gripper 单位：m，约 0=闭合、0.08=张开
时间戳：ROS header timestamp；MCAP 同时保留 log/publish timestamp

注意：/master/joint_* 是指令回显，不是 ROS2 控制入口；这正适合作为“实际下发并被控制器接收的目标量”。

State

关节/夹爪反馈：/puppet/joint_left、/puppet/joint_right，约 200 Hz
14D 顺序与 action 相同：[L_j0..L_j5, R_j0..R_j5, L_gripper, R_gripper]
单位：关节 rad，gripper m
末端位姿反馈：/puppet/end_pose_left、/puppet/end_pose_right
每臂末端位姿：[x, y, z, qx, qy, qz, qw]，位置单位 m

Metadata

instruction、episode 起止时间、FPS、topic 频率和消息数在 *_info.json 与任务级 info.json 中。
机器人型号：Cobot Magic；制作咖啡、制作三明治、开关台灯的任务级 metadata 明确记录，毛巾样本由相同 topic/TF/设备数据模式确认。
4 条样本均来自任务正常目录而不是 abnormal/；但源数据没有逐 episode 的显式布尔 success 字段，因此 manifest.json 中保留为 null。
源 metadata 未提供急停、人工接管、控制中断的逐 episode 标记；不能据空缺推断为“未发生”。

校验

OSS 下载命令全部成功，文件尺寸与远程对象完全一致。
使用 MCAP reader 对 4 个文件全量顺序读取并启用 chunk CRC 校验。
全量读取消息数分别与 metadata.yaml / *_info.json 一致：312,795、187,742、169,819、119,701。
所需三路 RGB、左右 action/state 和左右末端位姿在全部样本中均非空。

当前交付保留原始 MCAP，不额外生成 MP4/NPY，以避免无必要的重编码或同步重采样。如果后续需要文档推荐的 episode_xxxxxx/{head.mp4, action.npy, ...} 结构，应从 MCAP 按各自原始 timestamp 无损导出，并保留本目录作为不可变原始来源。

Downloads last month: 799

Total file size:

322 GB