fix(curriculum): Make boundaries in curriculum more sensible (#407)

* init

* fix tests

* unify codeio

* filtered for libraries not present in reasoning-gym

* fix more bounds

* puzzle24

* knight swap curriculum

* fix number sorting

* fix attributes

* add validation of config in creation of dataset

* dry run for instantiating and validating the datasets

* remove unused imports

* fix curriculum tests to reference newly updated attribute names
This commit is contained in:
Zafir Stojanovski 2025-04-04 20:24:14 +02:00 committed by GitHub
parent 7853263650
commit dced3bfc45
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
132 changed files with 1226 additions and 347 deletions

View file

@ -7,7 +7,7 @@ from typing import Any, Optional
import numpy as np
from ..coaching import BaseCurriculum, RangeAttributeDefinition
from ..coaching import BaseCurriculum, RangeAttributeDefinition, ScalarAttributeDefinition
from ..factory import ProceduralDataset, register_dataset
DATASET_NAME = "number_sorting"
@ -170,7 +170,7 @@ class NumberSortingCurriculum(BaseCurriculum):
self._define_attributes(
RangeAttributeDefinition(
name="numbers",
levels=list(range(5, 20, 2)),
levels=[10, 50, 100, 200],
description="How many numbers to sort",
lower_field_name="min_numbers",
upper_field_name="max_numbers",
@ -184,13 +184,17 @@ class NumberSortingCurriculum(BaseCurriculum):
upper_field_name="max_decimals",
ensure_interval=True,
),
RangeAttributeDefinition(
name="value",
levels=[-10_000, 10_000],
description="Range of numbers to sort",
lower_field_name="min_value",
upper_field_name="max_value",
ensure_interval=True,
ScalarAttributeDefinition(
name="min_value",
field_name="min_value",
levels=[-100, -500, -1000, -10000],
description="Minimum number value",
),
ScalarAttributeDefinition(
name="max_value",
field_name="max_value",
levels=[100, 500, 1000, 10000],
description="Maximum number value",
),
)