Datasets:

Yifei2vec
/

canny

Modalities:

Dataset Preview

Split (2)

train

The full dataset viewer is not available (click to read why). Only showing a preview of the rows.

The dataset generation failed

Error code:   DatasetGenerationError
Exception:    CastError
Message:      Couldn't cast
status: string
tasks: list<item: string>
  child 0, item: string
episodes: int64
action_frames: int64
video_files: int64
verified_video_frames: int64
removed_unreferenced_video_frames: int64
all_video_frames_referenced: bool
no_other_task_frames: bool
all_decoded_rgb_pixels_exact: bool
all_pts_verified: bool
data_and_statistics_unchanged: bool
input_unchanged: bool
encoded_bytes: int64
elapsed_seconds: double
all_stored_video_frames_referenced: bool
source_rgb_pixels_exact: bool
decoded_training_rgb_canny_pairs: int64
no_gaps_overlaps_or_extra_video_files: bool
task_names: list<item: string>
  child 0, item: string
training_episodes: int64
recomputed_canny_pixels_exact: bool
referenced_video_frames: int64
to
{'status': Value('string'), 'task_names': List(Value('string')), 'training_episodes': Value('int64'), 'video_files': Value('int64'), 'all_stored_video_frames_referenced': Value('bool'), 'no_gaps_overlaps_or_extra_video_files': Value('bool'), 'referenced_video_frames': Value('int64'), 'decoded_training_rgb_canny_pairs': Value('int64'), 'source_rgb_pixels_exact': Value('bool'), 'recomputed_canny_pixels_exact': Value('bool')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1827, in _prepare_split_single
                  for key, table in generator:
                                    ^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 613, in wrapped
                  for item in generator(*args, **kwargs):
                              ~~~~~~~~~^^^^^^^^^^^^^^^^^
                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
              tasks: list<item: string>
                child 0, item: string
              episodes: int64
              action_frames: int64
              video_files: int64
              verified_video_frames: int64
              removed_unreferenced_video_frames: int64
              all_video_frames_referenced: bool
              no_other_task_frames: bool
              all_decoded_rgb_pixels_exact: bool
              all_pts_verified: bool
              data_and_statistics_unchanged: bool
              input_unchanged: bool
              encoded_bytes: int64
              elapsed_seconds: double
              all_stored_video_frames_referenced: bool
              source_rgb_pixels_exact: bool
              decoded_training_rgb_canny_pairs: int64
              no_gaps_overlaps_or_extra_video_files: bool
              task_names: list<item: string>
                child 0, item: string
              training_episodes: int64
              recomputed_canny_pixels_exact: bool
              referenced_video_frames: int64
              to
              {'status': Value('string'), 'task_names': List(Value('string')), 'training_episodes': Value('int64'), 'video_files': Value('int64'), 'all_stored_video_frames_referenced': Value('bool'), 'no_gaps_overlaps_or_extra_video_files': Value('bool'), 'referenced_video_frames': Value('int64'), 'decoded_training_rgb_canny_pairs': Value('int64'), 'source_rgb_pixels_exact': Value('bool'), 'recomputed_canny_pixels_exact': Value('bool')}
              because column names don't match
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1694, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1880, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

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.

status string	task_names list	training_episodes int64	video_files int64	all_stored_video_frames_referenced bool	no_gaps_overlaps_or_extra_video_files bool	referenced_video_frames int64	decoded_training_rgb_canny_pairs int64	source_rgb_pixels_exact bool	recomputed_canny_pixels_exact bool
passed	[ "arrange_largest_number", "fold_clothes", "hang_mugs", "make_toast", "pack_objects_into_box", "pour_liquid_into_cup", "push_T", "sort_nesting_dolls_by_size", "stack_blocks", "stack_bowls", "store_laptop_and_headphones", "sweep_blocks" ]	1,192	408	true	true	3,533,790	324	true	true

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

Check out the documentation for more information.

RoboDojo：仅保留有 random 扩展的 12 个任务

状态：已完成。408 个视频文件、3,533,790 帧通过像素及时间戳核验；全部文件无未引用帧，324 对训练样本的独立读取验证通过。

范围

本目录只包含下列 12 个有 random 扩展的基础任务，共 1,192 条有效训练演示、588,965 个时间步。

这里是基础任务的标准训练演示，不是 _random 评测变体的演示。 这是用户给定源数据所提供的内容。任务筛选来自官方 leaderboard 与本地 *_random.yml 配置的匹配，证据保存在 meta/references/。

任务	训练演示	时间步
`arrange_largest_number`	98	57,138
`fold_clothes`	99	30,049
`hang_mugs`	100	52,853
`make_toast`	98	66,077
`pack_objects_into_box`	100	83,422
`pour_liquid_into_cup`	100	24,276
`push_T`	100	32,304
`sort_nesting_dolls_by_size`	100	59,013
`stack_blocks`	98	26,033
`stack_bowls`	100	44,874
`store_laptop_and_headphones`	100	52,734
`sweep_blocks`	99	60,192

其他任务不进入训练索引，物理视频中也不保留其他任务的边界片段。每个视频文件中的每一帧都被本目录的训练 metadata 引用，无闲置帧、重复引用或区间空隙。

源数据中相机缺帧的 8 条演示（源编号 14, 75, 959, 1520, 1576, 2939, 2945, 3456）沿用之前的排除方式。本目录仅保留 1,192 条有效演示的视频，其余帧不保留。

增强效果与训练读取

沿用已经调整好的数字细节版本：细尺度 RGB Canny 15/40、Gaussian σ=0.5，轻度双边去噪；粗尺度轮廓用于抑制背景纹理，并保护小物体内部的数字孔洞。不重新调整 Canny 参数。

数字细节对比：原图／上一版／当前细节版。
顶部布局视频、腕部抓取视频。
原生分辨率 640×480、25 FPS，三路 RGB 加三路 Canny。

RGB	Canny
`observation.images.cam_high`	`observation.images.cam_high_canny`
`observation.images.cam_left_wrist`	`observation.images.cam_left_wrist_canny`
`observation.images.cam_right_wrist`	`observation.images.cam_right_wrist_canny`

将本目录设为 LeRobot v3 数据根目录即可。所有相机视频统一为 H.264 crf=0 无损编码，输出解码后的 RGB 像素与整理前保留的对应帧完全相同。视频时间区间已重排，RGB/Canny 同步更新；episode 内 timestamp、frame_index、动作、状态和图像统计保持不变。

背景处理是纹理抑制，不是严格语义分割；少量墙边、桌沿仍可能保留，极小数字孔洞受原图分辨率限制。

校验与追溯

meta/packing_plan.json：输入快照和“仅保留训练帧”的规则。
meta/packing_receipts/：逐文件保留帧区间、原始演示编号、输出像素 SHA-256 与时间戳验证。
meta/validation_report.json：全视频、全帧的像素与时间戳核验。
meta/training_read_validation.json：独立检查全部 408 个文件中的帧覆盖范围，并抽取 324 对训练样本，从原始数据重算 Canny 核对。
meta/source_mapping.json：训练演示与原始演示的对应关系。
meta/all_converted_episodes.jsonl：本目录实际保留的 1,192 条演示及其新视频区间。
meta/source_alignment_issues.json：排除的原始缺帧演示记录。
meta/canny_plan.json：原始 Canny 参数与任务筛选证据；任务条目中的原始总量是筛选前源数据数量，实际保留量见 retained_training_episodes 和 retained_training_frames。

预览画面来自这 12 个任务。无需依赖上一版数据集的符号链接读取训练视频；部分无需裁剪的 Canny 文件使用硬链接复用存储。

复现与续跑

cd /home/yifei/RoboDojo
OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 \
/home/yifei/envs/do_wam/bin/python scripts/RoboDojo/pack_random_only_canny.py \
  --input data/RoboDojo_ee_lerobot_v30_video_detail_canny \
  --output data/RoboDojo_ee_lerobot_v30_video_random_only_detail_canny \
  --workers 128

OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 \
/home/yifei/envs/do_wam/bin/python scripts/RoboDojo/verify_random_only_canny.py \
  --root data/RoboDojo_ee_lerobot_v30_video_random_only_detail_canny

Downloads last month: 200

Total file size:

168 GB