Dataset Viewer
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
run: string
command_vx: double
episodes: int64
seconds: double
randomized: bool
pushes: bool
forward_speed_error: double
mean_forward_speed: double
fall_rate: double
mean_time_up_fraction: double
mean_spin_rad_s: double
mean_turns: double
drop_rate: double
mean_time_held_fraction: double
mean_rotation_rad: double
to
{'run': Value('string'), 'episodes': Value('int64'), 'seconds': Value('float64'), 'randomized': Value('bool'), 'mean_spin_rad_s': Value('float64'), 'mean_rotation_rad': Value('float64'), 'mean_turns': Value('float64'), 'drop_rate': Value('float64'), 'mean_time_held_fraction': Value('float64')}
because column names don't match
Traceback: Traceback (most recent call last):
File "/src/services/worker/src/worker/utils.py", line 147, 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 127, 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 483, 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 2840, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2373, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2398, 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 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
run: string
command_vx: double
episodes: int64
seconds: double
randomized: bool
pushes: bool
forward_speed_error: double
mean_forward_speed: double
fall_rate: double
mean_time_up_fraction: double
mean_spin_rad_s: double
mean_turns: double
drop_rate: double
mean_time_held_fraction: double
mean_rotation_rad: double
to
{'run': Value('string'), 'episodes': Value('int64'), 'seconds': Value('float64'), 'randomized': Value('bool'), 'mean_spin_rad_s': Value('float64'), 'mean_rotation_rad': Value('float64'), 'mean_turns': Value('float64'), 'drop_rate': Value('float64'), 'mean_time_held_fraction': Value('float64')}
because column names don't matchNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
Humanoid lab experiments
Every experiment from the first campaign of a home humanoid-robot project (September 2026): a simulated ROBOTIS OP3 humanoid that walks and a LEAP hand that spins a cube, trained with reinforcement learning on a home server's CPU and tested the way a real robot would be. Each experiment's pass bar was written before it ran, and failures are published as they came out. Simulation only.
The campaign's progress in 53 seconds.
Experiments
| experiment | what | verdict |
|---|---|---|
| 01-hand-spin | 01 · Robot hand spins a cube | fail |
| 02-hardened-walking | 02 · Hardened walking | fail |
| 03-robustness-matrix | 03 · Robustness matrix | fail |
| 04-walking-mild-pushes | 04-walking-mild-pushes | crashed |
| 05-walking-stronger-pushes | 05 · Walking with stronger shoves | fail |
| 07-hand-brain-demo | 07 · The hand obeys plain language | pass |
Each folder has README.md (the result against the bar and how it ran) and result.json; depending on
the experiment also videos, contact sheets, evaluations (eval-*.json), speed sweeps, gesture checks
and robustness matrices (robustness.json, robustness.md).
What was learned
- The baseline walking policy survives servo delay, stiffness errors, extra mass, a slippery floor and sensor errors with 0 % falls, but any shove from 30 N knocks it over.
- Training with strong kicks (1–5 m/s every 1–3 s) did not learn: 80.5 % falls, flat reward.
- Gentler, staged shoves cut falls in the training simulator to 20.3 %, but on the robot's control loop the policy still falls at 30 N, and it walks in a narrower speed range.
- The hand drops the cube in 2.3 % of randomized tests and spins at 0.94 rad/s (goal 1.0).
- The hand follows a plain-language instruction through a locally running language model (pass).
- Experiment 04 crashed: its training ran out of memory on a shared server.
Models
- Walking: YauhenBichel/op3-humanoid-walking-ppo
- Hand: YauhenBichel/leap-hand-cube-spin-ppo
- Code for the robot, control loop and simulator: humanoid-companion
Paths inside the reports (runs/…, exports/…, reports/…) refer to the project's working tree; the
videos and numbers are all here. Not a medical or care device; not a safety-rated controller.
- Downloads last month
- 59