Farabi Embodied Intelligence
Collection
Multilingual speech, multimodal perception, language-guided robotics, and compact embodied-AI models, datasets, and reproducibility artifacts. • 6 items • Updated
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
schema: string
runs: struct<smolvla_native_confirm: struct<summary: struct<n_records: int64, n_completed: int64, technica (... 267099 chars omitted)
child 0, smolvla_native_confirm: struct<summary: struct<n_records: int64, n_completed: int64, technical_errors: list<item: null>, sco (... 25059 chars omitted)
child 0, summary: struct<n_records: int64, n_completed: int64, technical_errors: list<item: null>, scoring: string, ho (... 3102 chars omitted)
child 0, n_records: int64
child 1, n_completed: int64
child 2, technical_errors: list<item: null>
child 0, item: null
child 3, scoring: string
child 4, horizons: struct<60: struct<success: int64, n: int64, fraction: double, exact_95_ci: list<item: double>, succe (... 351 chars omitted)
child 0, 60: struct<success: int64, n: int64, fraction: double, exact_95_ci: list<item: double>, success_per_atte (... 116 chars omitted)
child 0, success: int64
child 1, n: int64
child 2, fraction: double
child 3, exact_95_ci: list<item: double>
child 0, item: double
child 4, success_per_attempt: double
child 5, attempt_exact_95_ci: list<item: double>
child 0, item: double
child 6, attempts: int64
child 7, alternative: null
child 8, both: null
c
...
child 4, chunk_size: int64
child 5, n_action_steps: int64
child 6, n_obs_steps: int64
child 7, resize_imgs_with_padding: list<item: int64>
child 0, item: int64
child 8, input_features: struct<observation.state: list<item: int64>, observation.images.camera1: list<item: int64>, observat (... 85 chars omitted)
child 0, observation.state: list<item: int64>
child 0, item: int64
child 1, observation.images.camera1: list<item: int64>
child 0, item: int64
child 2, observation.images.camera2: list<item: int64>
child 0, item: int64
child 3, observation.images.camera3: list<item: int64>
child 0, item: int64
child 9, output_features: struct<action: list<item: int64>>
child 0, action: list<item: int64>
child 0, item: int64
child 2, git_commit: string
child 3, runner_sha256: string
child 4, packages: struct<numpy: string, mujoco: string, dm-control: string>
child 0, numpy: string
child 1, mujoco: string
child 2, dm-control: string
child 15, result_hashes: struct<select_fruit_1001: string>
child 0, select_fruit_1001: string
extension.json: string
README.md: string
reproduction.zip: string
confirmation.json: string
to
{'README.md': Value('string'), 'confirmation.json': Value('string'), 'extension.json': Value('string'), 'reproduction.zip': Value('string')}
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
schema: string
runs: struct<smolvla_native_confirm: struct<summary: struct<n_records: int64, n_completed: int64, technica (... 267099 chars omitted)
child 0, smolvla_native_confirm: struct<summary: struct<n_records: int64, n_completed: int64, technical_errors: list<item: null>, sco (... 25059 chars omitted)
child 0, summary: struct<n_records: int64, n_completed: int64, technical_errors: list<item: null>, scoring: string, ho (... 3102 chars omitted)
child 0, n_records: int64
child 1, n_completed: int64
child 2, technical_errors: list<item: null>
child 0, item: null
child 3, scoring: string
child 4, horizons: struct<60: struct<success: int64, n: int64, fraction: double, exact_95_ci: list<item: double>, succe (... 351 chars omitted)
child 0, 60: struct<success: int64, n: int64, fraction: double, exact_95_ci: list<item: double>, success_per_atte (... 116 chars omitted)
child 0, success: int64
child 1, n: int64
child 2, fraction: double
child 3, exact_95_ci: list<item: double>
child 0, item: double
child 4, success_per_attempt: double
child 5, attempt_exact_95_ci: list<item: double>
child 0, item: double
child 6, attempts: int64
child 7, alternative: null
child 8, both: null
c
...
child 4, chunk_size: int64
child 5, n_action_steps: int64
child 6, n_obs_steps: int64
child 7, resize_imgs_with_padding: list<item: int64>
child 0, item: int64
child 8, input_features: struct<observation.state: list<item: int64>, observation.images.camera1: list<item: int64>, observat (... 85 chars omitted)
child 0, observation.state: list<item: int64>
child 0, item: int64
child 1, observation.images.camera1: list<item: int64>
child 0, item: int64
child 2, observation.images.camera2: list<item: int64>
child 0, item: int64
child 3, observation.images.camera3: list<item: int64>
child 0, item: int64
child 9, output_features: struct<action: list<item: int64>>
child 0, action: list<item: int64>
child 0, item: int64
child 2, git_commit: string
child 3, runner_sha256: string
child 4, packages: struct<numpy: string, mujoco: string, dm-control: string>
child 0, numpy: string
child 1, mujoco: string
child 2, dm-control: string
child 15, result_hashes: struct<select_fruit_1001: string>
child 0, select_fruit_1001: string
extension.json: string
README.md: string
reproduction.zip: string
confirmation.json: string
to
{'README.md': Value('string'), 'confirmation.json': Value('string'), 'extension.json': Value('string'), 'reproduction.zip': Value('string')}
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.
Robot Execution Validation
VLABench execution records for released SmolVLA and Pi0 policies: 320 primary confirmation attempts, a prospectively specified 60-attempt family-matched native validation extension, and 37 separately reported development attempts. Each native trajectory is evaluated at 60 and 200 actions. Paired instructions use the same initial state and requested-versus-alternative goal predicates.
| File | Contents |
|---|---|
confirmation.json |
Frozen primary results, per-trial measurements, summaries, and raw-file hashes |
extension.json |
Frozen validation-extension results and the prospective configuration selection |
reproduction.zip |
Evaluation/analysis code, tests, protocols, selected inputs, complete action traces, initial states, scene images, and environment versions |
SHA256SUMS.json |
File checksums |
Reproduce the analysis
Download the files from one fixed repository revision and extract reproduction.zip.
The archive has its own file-level SHA256SUMS.json. Python 3.10 or later with NumPy,
SciPy, Matplotlib, and pytest is sufficient for analysis; no GPU is required.
Use the recorded package versions for exact regeneration.
From the extracted directory:
cp /path/to/download/confirmation.json code/paper/reproduce/revision_results.json
cp /path/to/download/extension.json code/paper/reproduce/validation_extension_results.json
cd code
export PYTHONPATH="$PWD/src:$PWD"
python -m pytest tests -q
python -m experiments.revision.freeze --root ../confirmation --out ../confirmation_check.json
python -m experiments.revision.freeze_extension --root ../extension --out ../extension_check.json
python -m paper.reproduce.revision_assets --out ../latex_assets
python -m paper.reproduce.extension_assets --out ../latex_assets
The checked exports reproduce the downloaded JSON files. The generated assets include the original four-family native results, paired outcomes, progress, development results, family-matched extension and synthesis, and stopping-rule sensitivity. Technical errors remain distinct from observed goal outcomes. The paired first-goal analysis is post hoc; the original native-stop outcomes are retained alongside it.
Extension validation compares the frozen selected configurations with the recorded inputs, undoing only VLABench's automatic expansion of relative XML asset paths. No raw record or prospective selection is changed. The extension retains three Pi0 numerical failures.
Rerun execution
Use the package inventories in environments/versions.json and install these pinned
upstream sources and their assets in separate simulator and policy environments:
- VLABench, revision
cf588fe60c0c7282174fe979f5913170cfe69017. - SmolVLA checkpoint, revision
4fd586e12dc14b04d9d606ddbb77448df4f0ff29. - Pi0 checkpoint, revision
1ad73753a74d5cd97e67856664350f3f0baa21dc. - Author OpenPI implementation, revision
4483d1da6332da44115fe530e4e6fdd89bd57b13; use the includedopenpi_uv.lock.
Initialize a local Git checkout of code/ before execution so the evaluator can
record the reproduction's source revision. From code/, relocate archived asset
references into a fresh directory:
python -m experiments.revision.prepare_reproduction --inputs ../inputs \
--vlabench-root /path/to/VLABench/VLABench --out ../relocated
Set PROJECT_STORAGE_ROOT to the absolute path of ../relocated, VLABENCH_ROOT
to the installed benchmark package, and PYTHONPATH to code/src:code using absolute
paths. Set MUJOCO_GL=egl, OMP_NUM_THREADS=4, and OPENBLAS_NUM_THREADS=4.
The evaluator permits physical GPUs 0 through 4 and enforces one renderer at a time.
Keep inference and rendering on separate GPUs.
Start the services in their respective policy environments:
python -m experiments.e2a.smolvla_server --gpu 1 --port 5581
python -m experiments.revision.pi0_server --gpu 3 --port 5583 \
--repo /path/to/openpi --checkpoint /path/to/pi0-checkpoint
Run the evaluator in the simulator environment with CUDA_VISIBLE_DEVICES=0.
For the extension, use --tasks select_book,select_mahjong,select_poker and the
relocated published track. For the original native control, use
--tasks select_fruit,select_toy,select_chemistry_tube,select_drink.
python -m experiments.revision.vlabench_control --out /path/to/fresh/run \
--tasks select_book,select_mahjong,select_poker \
--track-config ../relocated/track_1_in_distribution.json --episodes-per-task 10 \
--policy-seed 17 --max-substeps 1 --max-steps 200 --port 5581 --chunk 50
Use --port 5583 --chunk 5 for Pi0. For paired instructions, replace --tasks and
--track-config with --paired-spec ../relocated/final_study_command_spec.json.
Use fresh output directories and compare actual initial-state and observation hashes.
The common 200-action diagnostic budget is not every task's official leaderboard budget.
The separate synthetic-speech records are available in multilingual-robot-grounding-causal-audit.
- Downloads last month
- 24