mirror of
https://github.com/open-thought/reasoning-gym.git
synced 2026-04-19 12:58:07 +00:00
d0ef136d5b
* added curriculum * readapted readme * corrected small errors * Delete eval/eval/r1/algorithmic/word_sorting.json * removed redundant argument * added spell * removed duplicated fit * changed config * added composite changes * added composite changes * updated yaml * added spell backward * updated read me * added qwen2.5 * added * Add files via upload * updated missing trainer func * updated curr * updated spell back * updated correctness score func * updated configs * added local evals * added updates * updated datasets * added fsdp to hf utility * added algorithmic qwen 3b yaml * updated read me * updated configs * added preappend token * updated with thinking token * updated test score board * resolved comments * added evaluation scripts * removed results from pr * added config * added partial reward scoring * added evaluation composites * added training configs * added games eval * added rubriks cube * resolved merge cinflicts * added games config * added latest eval configs * updated strucutre * Delete training/evaluations/eval_graphs_composite.yaml --------- Co-authored-by: joesharratt1229 <joesharrat1229@gmail.com>
88 lines
3.3 KiB
Python
88 lines
3.3 KiB
Python
import pytest
|
|
|
|
from reasoning_gym.cognition.rubiks_cube import RubiksCubeConfig, RubiksCubeCurriculum, RubiksCubeDataset
|
|
|
|
|
|
def test_rubikscube_config_validation():
|
|
"""Test that invalid configs raise appropriate errors"""
|
|
with pytest.raises(AssertionError):
|
|
config = RubiksCubeConfig(cube_size=1) # Too small
|
|
config.validate()
|
|
|
|
with pytest.raises(AssertionError):
|
|
config = RubiksCubeConfig(max_scramble_steps=0) # Don't give an unscrambled cube
|
|
config.validate()
|
|
|
|
|
|
def test_rubikscube_deterministic():
|
|
"""Test that dataset generates same items with same seed"""
|
|
config = RubiksCubeConfig(seed=42, size=15) # Only check first 15 entries for speed
|
|
dataset1 = RubiksCubeDataset(config)
|
|
dataset2 = RubiksCubeDataset(config)
|
|
assert len(dataset1) == 15
|
|
assert len(dataset2) == 15
|
|
|
|
for i in range(len(dataset1)):
|
|
assert dataset1[i] == dataset2[i]
|
|
|
|
|
|
def test_rubikscube_items():
|
|
"""Test basic properties and solution of generated items"""
|
|
config = RubiksCubeConfig(
|
|
cube_size=3,
|
|
min_scramble_steps=4,
|
|
max_scramble_steps=4,
|
|
seed=42,
|
|
size=100,
|
|
)
|
|
dataset = RubiksCubeDataset(config)
|
|
|
|
for item in dataset:
|
|
assert isinstance(item, dict)
|
|
assert "question" in item
|
|
assert "answer" in item
|
|
assert "metadata" in item
|
|
|
|
# Check metadata contains required fields
|
|
assert "cube_size" in item["metadata"]
|
|
assert "cube_size" in item["metadata"]
|
|
assert "scramble_steps" in item["metadata"]
|
|
assert "scramble_moves" in item["metadata"]
|
|
assert "example_correct_answer" in item["metadata"]
|
|
|
|
assert dataset.score_answer(answer=item["metadata"]["example_correct_answer"], entry=item) == 1.0
|
|
assert dataset.score_answer(answer="a wrong solution", entry=item) == 0.01
|
|
assert dataset.score_answer(answer=None, entry=item) == 0.0
|
|
|
|
if item["metadata"]["example_correct_answer"] != "R":
|
|
assert dataset.score_answer(answer="R", entry=item) == 0.01
|
|
|
|
assert dataset.score_answer(answer="R2 R3 R4 R5 R'2 R'3", entry=item) == 0.01
|
|
|
|
if len(item["metadata"]["example_correct_answer"]) > 0:
|
|
assert dataset.score_answer(answer="", entry=item) == 0.01
|
|
|
|
|
|
def test_rubiks_cube_curriculum():
|
|
curriculum = RubiksCubeCurriculum()
|
|
|
|
base_value = {"size": 150, "seed": 1}
|
|
|
|
base_cfg: RubiksCubeConfig = curriculum.generate_configuration(base_value)
|
|
assert base_cfg.seed == 1
|
|
assert base_cfg.size == 150
|
|
assert base_cfg.cube_size == 3
|
|
assert base_cfg.min_scramble_steps == 3 and base_cfg.max_scramble_steps == 10
|
|
|
|
# test incrementing attribute levels for cube_size & scramble_stepsd attributes
|
|
curriculum.increment_attr_level("cube_size")
|
|
curriculum.increment_attr_level("scramble_steps")
|
|
increased_cfg = curriculum.generate_configuration(base_value)
|
|
assert increased_cfg.cube_size == 4
|
|
assert increased_cfg.min_scramble_steps == 3 and increased_cfg.max_scramble_steps == 25
|
|
|
|
# test decrementing attribute level for cube_size again
|
|
curriculum.decrement_attr_level("cube_size")
|
|
partially_decreased_cfg = curriculum.generate_configuration(base_value)
|
|
assert partially_decreased_cfg.cube_size == 3
|
|
assert partially_decreased_cfg.min_scramble_steps == 3 and partially_decreased_cfg.max_scramble_steps == 25
|