open-thought/reasoning-gym
1
0
Fork 0
mirror of https://github.com/open-thought/reasoning-gym.git synced 2026-04-19 12:58:07 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

updated algorithmics dataset (#269)

Browse source 
* updated algorithmic datasets
* added changes to symbolic and power
* updated power function test
This commit is contained in:
joesharratt1229 2025-03-05 23:32:53 +01:00 • committed by GitHub
parent f426db90ec
commit d9638df79c
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
5 changed files with 57 additions and 29 deletions
Download patch file Download diff file Expand all files Collapse all files

3
reasoning_gym/algorithmic/pool_matrix.py
View file

@ -13,6 +13,7 @@ The stride is equal to the kernel size, meaning there is no overlap between the
Your output should be a matrix in the same format as the input matrix.
The output matrix is smaller than the input matrix when the kernel size is greater than 1, and its elements may be floating-point numbers.
Give elements in the output matrix correct to 2 decimal places.
Perform {pool_type} pooling on the following matrix with a kernel size of {pool_size}:
{matrix}
@ -87,7 +88,7 @@ class PoolMatrixDataset(ProceduralDataset):
try:
oracle_answer = np.loadtxt(entry["answer"].splitlines(), dtype=np.float32)
answer = np.loadtxt(answer.splitlines(), dtype=np.float32)
if oracle_answer.shape == answer.shape and np.allclose(oracle_answer, answer):
if oracle_answer.shape == answer.shape and np.allclose(oracle_answer, answer, rtol=1e-2):
reward = 1.0
elif oracle_answer.shape == answer.shape:
reward = 0.1