fix(envs): Add source dataset and index to metadata (#388)

* add source dataset and index to metadata

* fix typo

* fix coach class and its test
This commit is contained in:
Zafir Stojanovski 2025-03-20 12:12:14 +01:00 committed by GitHub
parent 7475a20700
commit ce0a6c4878
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
104 changed files with 549 additions and 146 deletions

View file

@ -14,6 +14,8 @@ from ..factory import ProceduralDataset, register_dataset
QUESTION_TEMPLATE = """Count how many prime numbers there are between {start} and {end} (inclusive) ?"""
DATASET_NAME = "count_primes"
@dataclass
class CountPrimesConfig:
@ -60,6 +62,8 @@ class CountPrimesDataset(ProceduralDataset):
"question": QUESTION_TEMPLATE.format(start=start, end=end),
"answer": str(answer),
"metadata": {
"source_dataset": DATASET_NAME,
"source_index": idx,
"start": start,
"end": end,
"primes": primes,
@ -88,4 +92,4 @@ class CountPrimesCurriculum(BaseCurriculum):
)
register_dataset("count_primes", CountPrimesDataset, CountPrimesConfig, CountPrimesCurriculum)
register_dataset(DATASET_NAME, CountPrimesDataset, CountPrimesConfig, CountPrimesCurriculum)