The dataset viewer is not available for this subset.
Exception: SplitsNotFoundError
Message: The split names could not be parsed from the dataset config.
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 286, in get_dataset_config_info
for split_generator in builder._split_generators(
~~~~~~~~~~~~~~~~~~~~~~~~~^
StreamingDownloadManager(base_path=builder.base_path, download_config=download_config)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/webdataset/webdataset.py", line 80, in _split_generators
raise ValueError(
...<2 lines>...
)
ValueError: The TAR archives of the dataset should be in WebDataset format, but the files in the archive don't share the same prefix or the same types.
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 68, in compute_split_names_from_streaming_response
for split in get_dataset_split_names(
~~~~~~~~~~~~~~~~~~~~~~~^
path=dataset,
^^^^^^^^^^^^^
config_name=config,
^^^^^^^^^^^^^^^^^^^
token=hf_token,
^^^^^^^^^^^^^^^
)
^
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 340, in get_dataset_split_names
info = get_dataset_config_info(
path,
...<6 lines>...
**config_kwargs,
)
File "/usr/local/lib/python3.14/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.
RoboCasa RLDX-1 entropy and CPU reconstruction kit
7200 complete episodes / 2073457 original observations. All entropy arrays and episode-safe slow2/fast4 action targets validated before publishing.
Source: kimtaey/robocasa_mg_gr00t_300. Checkpoint: prehj/rldx1-robocasa-img-gb64-60k-baseline.
No GPU, model checkpoint, RLDX installation or entropy re-inference is needed.
Saved precision labels are included; no clustering rerun is needed.
python -m pip install huggingface_hub
hf download prehj/robocasa-rldx-demospeedup-entropy --repo-type dataset --local-dir entropy-kit
cd entropy-kit
sha256sum -c SHA256SUMS
tar -xzf robocasa-entropy.tar.gz
python -m pip install -r robocasa-entropy/requirements.txt
python robocasa-entropy/code/rebuild.py --source /path/to/robocasa_mg_gr00t_300 --out /path/to/new-speedup-dataset
Python 3.10 was used. Original dataset including videos is required. Use a new output directory and allow space for a full dataset copy. Reconstruction is CPU-only and copies original videos. --verify-only verifies original parquet and entropy SHA256 without generating.
The output preserves original observations and actions and adds speedup.actions (reshape [8,12]), speedup.valid, source intervals, rates, and entropy. Train on these derived chunks, not the original action. Use generated meta/speedup_action_stats.json for target normalization. See robocasa-entropy/code/FORMAT.md for the complete reading contract.
Slow=2, fast=4. Sum raw arm deltas (dimensions5:11) and zero base deltas (0:4); preserve final control mode (4) and gripper (11). No clipping, SO(3) correction, or cross-episode merging. This is an observation-preserving delta-action adaptation of DemoSpeedup, not row subsampling. Provenance and protocol are bundled. Source parquet identity is checked; source video identity is not pinned by this small kit (the full release checks its copied videos).
- Downloads last month
- 36