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
Do the leaf sorting experiment. compound leaves go to the left side and simple leaves go to the righ (... 6 chars omitted): struct<train: list<item: string>, test: list<item: string>>
child 0, train: list<item: string>
child 0, item: string
child 1, test: list<item: string>
child 0, item: string
Do the leaf sorting experiment. compound leaves go to the right side and simple leaves go to the lef (... 6 chars omitted): struct<train: list<item: string>, test: list<item: string>>
child 0, train: list<item: string>
child 0, item: string
child 1, test: list<item: string>
child 0, item: string
Do the redox reaction experiment. lift the metal plate out of the left beaker: struct<train: list<item: string>, test: list<item: string>>
child 0, train: list<item: string>
child 0, item: string
child 1, test: list<item: string>
child 0, item: string
Do the redox reaction experiment. move the plate over the blue solution: struct<train: list<item: string>, test: list<item: string>>
child 0, train: list<item: string>
child 0, item: string
child 1, test: list<item: string>
child 0, item: string
Do the redox reaction experiment. point at the plate with the left hand: struct<train: list<item: string>, test: list<item: string>>
child 0, train: list<item: string>
child 0, item: string
child 1, test: list<item: string>
child 0, item: string
Do the redox reaction experiment. lower the plate into the blue solution: struct<t
...
child 0, item: int64
child 85, 283: list<item: list<item: int64>>
child 0, item: list<item: int64>
child 0, item: int64
child 86, 284: list<item: list<item: int64>>
child 0, item: list<item: int64>
child 0, item: int64
child 87, 285: list<item: list<item: int64>>
child 0, item: list<item: int64>
child 0, item: int64
child 88, 286: list<item: list<item: int64>>
child 0, item: list<item: int64>
child 0, item: int64
child 89, 287: list<item: list<item: int64>>
child 0, item: list<item: int64>
child 0, item: int64
child 90, 289: list<item: list<item: int64>>
child 0, item: list<item: int64>
child 0, item: int64
child 91, 290: list<item: list<item: int64>>
child 0, item: list<item: int64>
child 0, item: int64
sha256_of_episodes: string
frames_in_train: int64
rule: string
pre_frames: int64
only_from_episode: null
total_frames_in_these_episodes: int64
sentences: list<item: string>
child 0, item: string
base: struct<labels: string, labels_revision: string, merged_dataset: string, merged_revision: string, not (... 10 chars omitted)
child 0, labels: string
child 1, labels_revision: string
child 2, merged_dataset: string
child 3, merged_revision: string
child 4, note: string
range_unit: string
version: int64
made_with: string
joints: string
episodes_with_ranges: int64
threshold_deg: double
train_episodes: list<item: int64>
child 0, item: int64
to
{'version': Value('int64'), 'rule': Value('string'), 'description': Value('string'), 'sentences': List(Value('string')), 'threshold_deg': Value('float64'), 'joints': Value('string'), 'pre_frames': Value('int64'), 'only_from_episode': Value('null'), 'range_unit': Value('string'), 'train_episodes': List(Value('int64')), 'frames_in_train': Value('int64'), 'total_frames_in_these_episodes': Value('int64'), 'episodes_with_ranges': Value('int64'), 'episodes': {'181': List(List(Value('int64'))), '182': List(List(Value('int64'))), '183': List(List(Value('int64'))), '184': List(List(Value('int64'))), '185': List(List(Value('int64'))), '186': List(List(Value('int64'))), '187': List(List(Value('int64'))), '189': List(List(Value('int64'))), '190': List(List(Value('int64'))), '191': List(List(Value('int64'))), '192': List(List(Value('int64'))), '193': List(List(Value('int64'))), '194': List(List(Value('int64'))), '195': List(List(Value('int64'))), '196': List(List(Value('int64'))), '197': List(List(Value('int64'))), '198': List(List(Value('int64'))), '199': List(List(Value('int64'))), '200': List(List(Value('int64'))), '201': List(List(Value('int64'))), '202': List(List(Value('int64'))), '203': List(List(Value('int64'))), '204': List(List(Value('int64'))), '205': List(List(Value('int64'))), '206': List(List(Value('int64'))), '207': List(List(Value('int64'))), '209': List(List(Value('int64'))), '210': List(List(Value('int64'))), '211': List(List(Value('int64'))), '212': List(List(Value('int
...
, '240': List(List(Value('int64'))), '252': List(List(Value('int64'))), '254': List(List(Value('int64'))), '255': List(List(Value('int64'))), '256': List(List(Value('int64'))), '257': List(List(Value('int64'))), '258': List(List(Value('int64'))), '259': List(List(Value('int64'))), '260': List(List(Value('int64'))), '261': List(List(Value('int64'))), '262': List(List(Value('int64'))), '263': List(List(Value('int64'))), '264': List(List(Value('int64'))), '265': List(List(Value('int64'))), '266': List(List(Value('int64'))), '267': List(List(Value('int64'))), '269': List(List(Value('int64'))), '270': List(List(Value('int64'))), '271': List(List(Value('int64'))), '272': List(List(Value('int64'))), '273': List(List(Value('int64'))), '274': List(List(Value('int64'))), '276': List(List(Value('int64'))), '277': List(List(Value('int64'))), '278': List(List(Value('int64'))), '279': List(List(Value('int64'))), '280': List(List(Value('int64'))), '281': List(List(Value('int64'))), '282': List(List(Value('int64'))), '283': List(List(Value('int64'))), '284': List(List(Value('int64'))), '285': List(List(Value('int64'))), '286': List(List(Value('int64'))), '287': List(List(Value('int64'))), '289': List(List(Value('int64'))), '290': List(List(Value('int64')))}, 'sha256_of_episodes': Value('string'), 'made_with': Value('string'), 'base': {'labels': Value('string'), 'labels_revision': Value('string'), 'merged_dataset': Value('string'), 'merged_revision': Value('string'), 'note': 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 2951, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2461, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2486, 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 547, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 430, 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
Do the leaf sorting experiment. compound leaves go to the left side and simple leaves go to the righ (... 6 chars omitted): struct<train: list<item: string>, test: list<item: string>>
child 0, train: list<item: string>
child 0, item: string
child 1, test: list<item: string>
child 0, item: string
Do the leaf sorting experiment. compound leaves go to the right side and simple leaves go to the lef (... 6 chars omitted): struct<train: list<item: string>, test: list<item: string>>
child 0, train: list<item: string>
child 0, item: string
child 1, test: list<item: string>
child 0, item: string
Do the redox reaction experiment. lift the metal plate out of the left beaker: struct<train: list<item: string>, test: list<item: string>>
child 0, train: list<item: string>
child 0, item: string
child 1, test: list<item: string>
child 0, item: string
Do the redox reaction experiment. move the plate over the blue solution: struct<train: list<item: string>, test: list<item: string>>
child 0, train: list<item: string>
child 0, item: string
child 1, test: list<item: string>
child 0, item: string
Do the redox reaction experiment. point at the plate with the left hand: struct<train: list<item: string>, test: list<item: string>>
child 0, train: list<item: string>
child 0, item: string
child 1, test: list<item: string>
child 0, item: string
Do the redox reaction experiment. lower the plate into the blue solution: struct<t
...
child 0, item: int64
child 85, 283: list<item: list<item: int64>>
child 0, item: list<item: int64>
child 0, item: int64
child 86, 284: list<item: list<item: int64>>
child 0, item: list<item: int64>
child 0, item: int64
child 87, 285: list<item: list<item: int64>>
child 0, item: list<item: int64>
child 0, item: int64
child 88, 286: list<item: list<item: int64>>
child 0, item: list<item: int64>
child 0, item: int64
child 89, 287: list<item: list<item: int64>>
child 0, item: list<item: int64>
child 0, item: int64
child 90, 289: list<item: list<item: int64>>
child 0, item: list<item: int64>
child 0, item: int64
child 91, 290: list<item: list<item: int64>>
child 0, item: list<item: int64>
child 0, item: int64
sha256_of_episodes: string
frames_in_train: int64
rule: string
pre_frames: int64
only_from_episode: null
total_frames_in_these_episodes: int64
sentences: list<item: string>
child 0, item: string
base: struct<labels: string, labels_revision: string, merged_dataset: string, merged_revision: string, not (... 10 chars omitted)
child 0, labels: string
child 1, labels_revision: string
child 2, merged_dataset: string
child 3, merged_revision: string
child 4, note: string
range_unit: string
version: int64
made_with: string
joints: string
episodes_with_ranges: int64
threshold_deg: double
train_episodes: list<item: int64>
child 0, item: int64
to
{'version': Value('int64'), 'rule': Value('string'), 'description': Value('string'), 'sentences': List(Value('string')), 'threshold_deg': Value('float64'), 'joints': Value('string'), 'pre_frames': Value('int64'), 'only_from_episode': Value('null'), 'range_unit': Value('string'), 'train_episodes': List(Value('int64')), 'frames_in_train': Value('int64'), 'total_frames_in_these_episodes': Value('int64'), 'episodes_with_ranges': Value('int64'), 'episodes': {'181': List(List(Value('int64'))), '182': List(List(Value('int64'))), '183': List(List(Value('int64'))), '184': List(List(Value('int64'))), '185': List(List(Value('int64'))), '186': List(List(Value('int64'))), '187': List(List(Value('int64'))), '189': List(List(Value('int64'))), '190': List(List(Value('int64'))), '191': List(List(Value('int64'))), '192': List(List(Value('int64'))), '193': List(List(Value('int64'))), '194': List(List(Value('int64'))), '195': List(List(Value('int64'))), '196': List(List(Value('int64'))), '197': List(List(Value('int64'))), '198': List(List(Value('int64'))), '199': List(List(Value('int64'))), '200': List(List(Value('int64'))), '201': List(List(Value('int64'))), '202': List(List(Value('int64'))), '203': List(List(Value('int64'))), '204': List(List(Value('int64'))), '205': List(List(Value('int64'))), '206': List(List(Value('int64'))), '207': List(List(Value('int64'))), '209': List(List(Value('int64'))), '210': List(List(Value('int64'))), '211': List(List(Value('int64'))), '212': List(List(Value('int
...
, '240': List(List(Value('int64'))), '252': List(List(Value('int64'))), '254': List(List(Value('int64'))), '255': List(List(Value('int64'))), '256': List(List(Value('int64'))), '257': List(List(Value('int64'))), '258': List(List(Value('int64'))), '259': List(List(Value('int64'))), '260': List(List(Value('int64'))), '261': List(List(Value('int64'))), '262': List(List(Value('int64'))), '263': List(List(Value('int64'))), '264': List(List(Value('int64'))), '265': List(List(Value('int64'))), '266': List(List(Value('int64'))), '267': List(List(Value('int64'))), '269': List(List(Value('int64'))), '270': List(List(Value('int64'))), '271': List(List(Value('int64'))), '272': List(List(Value('int64'))), '273': List(List(Value('int64'))), '274': List(List(Value('int64'))), '276': List(List(Value('int64'))), '277': List(List(Value('int64'))), '278': List(List(Value('int64'))), '279': List(List(Value('int64'))), '280': List(List(Value('int64'))), '281': List(List(Value('int64'))), '282': List(List(Value('int64'))), '283': List(List(Value('int64'))), '284': List(List(Value('int64'))), '285': List(List(Value('int64'))), '286': List(List(Value('int64'))), '287': List(List(Value('int64'))), '289': List(List(Value('int64'))), '290': List(List(Value('int64')))}, 'sha256_of_episodes': Value('string'), 'made_with': Value('string'), 'base': {'labels': Value('string'), 'labels_revision': Value('string'), 'merged_dataset': Value('string'), 'merged_revision': Value('string'), 'note': 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.
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
SciEdu_text_aug
OpenArm SciEdu pi0.5 학습에 쓰는 같은 뜻 문장 목록이다. H200 학습 서버의 scripts/text_aug/를 그대로 올렸다.
형식
{"<학습 원문(데이터셋 meta/tasks.parquet의 문장)>": {"train": ["같은 뜻 문장 6개"], "test": ["학습에 넣지 않는 같은 뜻 문장 2개"]}}
train은 학습 때 쓴다.test는 평가 전용이고 학습에 넣지 않는다.- 앞부분("Do the ... experiment. ")은 대부분 그대로 두었다. 원문마다
train하나는 앞부분을 바꿨다(예: "Redox reaction experiment: ..."). - left/right, dipped, "then stop above it"처럼 문장을 가르는 말은 원문과 같게 두었다.
파일
| 파일 | 원문 | 맞는 데이터셋 |
|---|---|---|
paraphrases_sciedu_ctx_v2.json |
20개, 학습용 12 + 평가용 2 + 먼 평가용 2 | SciEdu_291epi. SciEdu_291epi_aug_pi05 학습에 쓴다 |
paraphrases_sciedu_ctx.json |
20개 (나뭇잎 방향 2 + Redox 9 + 검전기 7 + 맥락별 2) | SciEdu_pauses_3exp_210epi_leafdir_ctx, ..._ctx2와 같은 문장을 쓰는 데이터셋 |
paraphrases_pauses3_leafdir.json |
18개 (나뭇잎 방향 2 + Redox 9 + 검전기 7) | SciEdu_pauses_3exp_210epi_leafdir |
paraphrases_pauses3.json |
18개 (나뭇잎 규칙 2 + Redox 9 + 검전기 7) | SciEdu_pauses_3exp_210epi |
학습에 넣는 법
데이터셋은 바꾸지 않는다. 학습 런처(scripts/train_pi05.py의 install_text_augmentation)가 lerobot-train의 배치 반복자(lerobot.scripts.lerobot_train.cycle)를 감싼다.
배치의 task 문장마다 확률 TEXT_AUG_P(기본 0.5)로 train 목록에서 하나를 무작위로 골라 바꾼다(시드 TEXT_AUG_SEED, 기본 0).
lerobot 코드는 고치지 않고, 저장되는 전처리기도 그대로라 추론은 바뀌지 않는다.
TEXT_AUG=scripts/text_aug/paraphrases_sciedu_ctx.json TEXT_AUG_P=0.5 ...
10/4에 이 방식으로 돌린 학습(..._leafdir_textaug_pi05)은 12,798스텝에서 멈추고 버렸다.
5천 스텝까지 문장의 절반(80,262 / 160,032)이 바뀌었고, 목록에 없는 문장은 0이었다. 이 방식으로 끝까지 학습한 모델은 아직 없다.
v2: 더 다양한 표현 (2026-10-06)
운영자가 더 다양하게 넣자고 해서 build_v2.py로 만들었다. v1의 학습용 6개·평가용 2개는 그대로 두고 원문마다 다음을 더했다.
- 학습용 6개
- 사물 이름 바꾸기: metal strip, blue liquid, fur cloth, leaf card
- 앞부분 꼴 바꾸기:
In the ... experiment,,Redox experiment:,For the ... experiment, please,... task. - 공손한 말과 문장 구조 바꾸기
test_far2개: 학습에 한 번도 나오지 않는 앞부분 꼴(During the ...,We are doing ... Now,This is ... Next,)과 이름(sheet, container, pad)을 쓴다.test는 학습 표현과 가까운 평가이고,test_far는 먼 평가다.
문장을 가르는 말은 build_v2.py가 기계로 검사한다: left/right, dipped, "위에서 멈춤", left beaker / blue solution, 실험 이름. 학습 런처는 train만 읽는다.
- Downloads last month
- 63