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/packaged_modules/json/json.py", line 291, in _generate_tables
pa_table = paj.read_json(
io.BytesIO(batch), read_options=paj.ReadOptions(block_size=block_size)
)
File "pyarrow/_json.pyx", line 364, in pyarrow._json.read_json
File "pyarrow/error.pxi", line 155, in pyarrow.lib.pyarrow_internal_check_status
return check_status(status)
File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
raise convert_status(status)
pyarrow.lib.ArrowInvalid: JSON parse error: Column() changed from object to string in row 0
During handling of the above exception, another exception occurred:
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/json/json.py", line 101, in _split_generators
pa_table = next(iter(self._generate_tables(**splits[0].gen_kwargs, allow_full_read=False)))[1]
~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 306, in _generate_tables
batch = json_encode_fields_in_json_lines(original_batch, json_field_paths)
File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 111, in json_encode_fields_in_json_lines
examples = [ujson_loads(line) for line in original_batch.splitlines()]
~~~~~~~~~~~^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 20, in ujson_loads
return pd.io.json.ujson_loads(*args, **kwargs)
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
ValueError: Expected object or value
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.
SONIC R1 retargeted motion dataset
Paired 26-DOF Unitree R1 robot references and 24-joint SMPL human-motion conditioning, prepared for SONIC R1 training. The exact training corpus contains 129,778 motion pairs, split into 127,288 training and 2,490 validation motions. Original/mirrored motion pairs share a split.
Motion Data by Bones Studio. Human SMPL conditioning comes from the bones_seed_smpl release accompanying NVIDIA GEAR-SONIC. See source terms and attribution.
Publication status: check DATASET_STATUS.json. Use the dataset once its state is complete; while uploading, some archive files may be absent. Final verification is recorded in dataset-verification.json.
Contents
| Path | Description |
|---|---|
shard-0000.tar … shard-0506.tar |
507 original training archives; 39,470,684,160 bytes in total (36.76 GiB) |
manifest.json |
Exact original training manifest, split counts and per-archive SHA-256 checksums |
metadata/packaging_records/ |
Per-motion split, frame counts, retargeting provenance and original SMPL hashes |
metadata/validation_manifest.json |
Independent validation-only file inventory |
extract_dataset.py |
Archive checksum verification and selective extraction without unpickling files |
dataset-verification.json |
Full corpus archive, pairing, split and payload-size verification |
Every archive contains matching entries such as:
robot/train/<motion_key>.pkl
robot/validation/<motion_key>.pkl
smpl/<motion_key>.pkl
Each motion belongs to exactly one robot split and has exactly one matching SMPL file. Robot references are serialized with joblib compression; human SMPL files retain the original released bytes. R1 references are 30 Hz, while SMPL conditioning is 50 Hz. Equal motion names identify pairs; their frame counts need not be equal. Use the SONIC motion loader's time resampling rather than matching array indices directly.
Robot data include joint/body motion fields consumed by SONIC's R1 motion library. SMPL files include fps, pose_aa, smpl_joints and transl. These are real human SMPL conditions, not SOMA joint arrays or dummy SMPL values.
Download and extract
python -m pip install huggingface_hub
hf download Jay-Ye/SONIC_R1_retarget --repo-type dataset \
--local-dir "$HOME/datasets/SONIC_R1_retarget"
python "$HOME/datasets/SONIC_R1_retarget/extract_dataset.py" \
--root "$HOME/datasets/SONIC_R1_retarget" \
--output "$HOME/datasets/SONIC_R1_expanded"
The helper verifies every archive before extraction and checks that all motion files are paired and assigned to the expected split. It does not deserialize pickle data. Select --split train or --split validation, and optionally --kind robot or --kind smpl; defaults are all splits and both kinds. Extraction requires space for the expanded files in addition to the downloaded archives. Pin an HF commit with hf download --revision <commit> to reproduce an experiment.
Training paths
After extraction:
SONIC_R1_expanded/robot/train/
SONIC_R1_expanded/robot/validation/
SONIC_R1_expanded/smpl/
Use the R1 SONIC training implementation and its matching robot configuration. Set manager_env.commands.motion.motion_lib_cfg.motion_file to robot/train/ and smpl_motion_file to smpl/. Evaluation uses robot/validation/. The original SageMaker training input can also consume this repository's unchanged root manifest and 507 tar files.
Dataset identity: d44780e842fec7cdcba7a0513162e326339073d20dade2a73a812939cc5441b5.
The training source, retargeting sources and R1 MJCF hashes are preserved in the manifest and packaging records. No raw Bones-SEED BVH, intermediate CSV, model checkpoints, credentials or private cloud configuration are included. The corresponding trained policies are published separately at Jay-Ye/SONIC_R1.
- Downloads last month
- 181