Datasets:
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
problem: string
answer: string
solution: string
source_repo: string
dropped_rows: list<item: struct<line: int64, reason: string>>
child 0, item: struct<line: int64, reason: string>
child 0, line: int64
child 1, reason: string
kept_rows: int64
dropped: struct<missing_or_blank_answer: int64, duplicate_problem: int64, over_quota: int64>
child 0, missing_or_blank_answer: int64
child 1, duplicate_problem: int64
child 2, over_quota: int64
total_rows: int64
to
{'source_repo': Value('string'), 'total_rows': Value('int64'), 'kept_rows': Value('int64'), 'dropped': {'missing_or_blank_answer': Value('int64'), 'duplicate_problem': Value('int64'), 'over_quota': Value('int64')}, 'dropped_rows': List({'line': Value('int64'), 'reason': 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
problem: string
answer: string
solution: string
source_repo: string
dropped_rows: list<item: struct<line: int64, reason: string>>
child 0, item: struct<line: int64, reason: string>
child 0, line: int64
child 1, reason: string
kept_rows: int64
dropped: struct<missing_or_blank_answer: int64, duplicate_problem: int64, over_quota: int64>
child 0, missing_or_blank_answer: int64
child 1, duplicate_problem: int64
child 2, over_quota: int64
total_rows: int64
to
{'source_repo': Value('string'), 'total_rows': Value('int64'), 'kept_rows': Value('int64'), 'dropped': {'missing_or_blank_answer': Value('int64'), 'duplicate_problem': Value('int64'), 'over_quota': Value('int64')}, 'dropped_rows': List({'line': Value('int64'), 'reason': 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.
DeepScaleR Curated (8,000 problems)
A quality-screened, 8,000-problem subset of the DeepScaleR math reasoning corpus, prepared as a drop-in training set for a Verl-based GRPO run reproducing the DeepSeek-R1 "aha moment" experiment on a small (1.5B) pretrained model.
Source
This dataset was filtered from agentica-org/DeepScaleR-Preview-Dataset
(repo id: agentica-org/DeepScaleR-Preview-Dataset), the training corpus open-sourced by the
DeepScaleR team (AIME 1984-2023, AMC pre-2023, Omni-MATH and Still).
What's in here
- 8,000 kept records out of 40,315 examined in the raw data file, taken as the earliest survivors in raw-file order.
- Every record keeps the original schema:
problem,answer,solution. - Row order and
extra_info.index(in the Verl parquet companion file) preserve the record's 0-based line number in the source file, so any row can be traced back to its origin.
Screening rules
| Rule | Dropped |
|---|---|
answer absent / null / empty / whitespace-only (missing_or_blank_answer) |
6 |
problem is a character-for-character repeat of an earlier record (duplicate_problem) |
928 |
beyond the 8,000-record training budget (over_quota) |
31,381 |
| kept | 8,000 |
The full per-record log of the quality drops lives in audit_report.json.
Files
deepscaler_curated.jsonl- the 8,000 kept records, one{"problem", "answer", "solution"}object per line, in raw-file order.audit_report.json- the screening paperwork (counts + per-record drop log).
Usage
from datasets import load_dataset
ds = load_dataset("dusersad12/deepscaler-curated", split="train")
For Verl, convert to the standard RL parquet layout (data_source, prompt, ability,
reward_model, extra_info) using extra_info.index as the provenance index.
License
MIT, inherited from the source dataset.
- Downloads last month
- 24