JHeisler/aloha_solo_left_act_removal_reversed_13k
Robotics • Updated • 39
Dataset Viewer
The dataset viewer is not available for this subset.
Cannot get the split names for the config 'default' of the dataset.
Exception: SplitsNotFoundError
Message: The split names could not be parsed from the dataset config.
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.12/site-packages/datasets/inspect.py", line 286, in get_dataset_config_info
for split_generator in builder._split_generators(
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/parquet/parquet.py", line 118, in _split_generators
self.info.features = datasets.Features.from_arrow_schema(pq.read_schema(f))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/features/features.py", line 1945, in from_arrow_schema
metadata_features = Features.from_dict(metadata["info"]["features"])
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/features/features.py", line 1983, in from_dict
obj = generate_from_dict(dic)
^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/features/features.py", line 1564, in generate_from_dict
return {key: generate_from_dict(value) for key, value in obj.items()}
^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/features/features.py", line 1570, in generate_from_dict
raise ValueError(f"Feature type '{_type}' not found. Available feature types: {list(_FEATURE_TYPES.keys())}")
ValueError: Feature type 'VideoFrame' not found. Available feature types: ['Value', 'ClassLabel', 'Translation', 'TranslationVariableLanguages', 'LargeList', 'List', 'Array2D', 'Array3D', 'Array4D', 'Array5D', 'Audio', 'Image', 'Video', 'Pdf', 'Nifti', 'Json']
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/split_names.py", line 65, in compute_split_names_from_streaming_response
for split in get_dataset_split_names(
^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/inspect.py", line 340, in get_dataset_split_names
info = get_dataset_config_info(
^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/inspect.py", line 291, in get_dataset_config_info
raise SplitsNotFoundError("The split names could not be parsed from the dataset config.") from err
datasets.inspect.SplitsNotFoundError: The split names could not be parsed from the dataset config.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.
aloha_solo_left_4_6_26_reversed
Synthetic mask-REMOVAL dataset derived by time-reversing every episode of JHeisler/aloha_solo_left_4_6_26.
Each placement episode (gripper opens, releasing mask onto face) becomes a removal episode (gripper closes, picking mask off face) when reversed. The action sequence is reversed AND shifted by 1 step so that new_action[k] correctly targets the previous state in forward time, which is the next state in reverse time.
Statistics
- Episodes: 50
- Total samples: 29735 (~1 fewer per episode than the source due to action-shift drop)
- fps: 30
- state_dim / action_dim: 9 / 9
- Cameras: cam_high, cam_left_wrist (3×480×640)
- Codec: HEVC (NVENC, cq=26, preset=p5) — re-encoded from source AV1 for faster generation. PyAV reads both transparently.
Caveats
- Visual transitions are physically backwards (mask appears on face at frame 0, gets pulled off). Doesn't affect ACT training (n_obs_steps=1, no temporal-coherence input), but matters for any human inspection.
- Gripper open/close events naturally invert (release ⇄ grasp), correctly matching removal semantics.
- Use this as a lower-bound baseline for mask-removal policy training. Native-removal-data trained policies will likely outperform.
Generation
See S005_ACT_Removal_Reversed/working/reverse_dataset.py in the project repo.
- Downloads last month
- 175