Commit graph

990 commits

Author SHA1 Message Date
rishabhranawat
2d57beb517 commit formatting 2025-02-10 22:05:45 -08:00
rishabhranawat
6e3d049fed [eval-v1] benchmark with 50 samples 2025-02-10 22:05:09 -08:00
rishabhranawat
06cabcfdee [eval-v1] add a simple readme with some details 2025-02-10 21:57:00 -08:00
rishabhranawat
615c63d2f9 [eval-v1] pre commit formatting 2025-02-10 21:50:22 -08:00
rishabhranawat
88c875c00f [eval-v1] add timer 2025-02-10 21:48:44 -08:00
rishabhranawat
be3d04e7cb [eval-v1] async to speed up inference/evaluation 2025-02-10 21:35:46 -08:00
joesharratt1229
a3ea4449d1 added r1 evaluation logic 2025-02-11 03:46:56 +00:00
tohskai
7bad77b426 Improve support for multivariate polynomials 2025-02-11 01:58:07 +01:00
Dragan Jovanović
55b0226ccf fix for isort 2025-02-11 00:20:46 +01:00
Dragan Jovanović
328d744780 initial draft for circuit_logic dataset generator 2025-02-11 00:09:00 +01:00
Andreas Koepf
eb25ab9656 update gallery, lower default config values for PowerFunctionDataset 2025-02-10 22:42:04 +01:00
Andreas Köpf
898dc0754a Merge pull request #100 from zafstojano/env/matrix-manipulation
Matrix Manipulation Dataset
2025-02-10 22:37:37 +01:00
Zafir Stojanovski
f255831f1c add more config params 2025-02-10 22:30:36 +01:00
Zafir Stojanovski
3e42d9588e count bits (#101) 2025-02-10 22:12:50 +01:00
Andreas Koepf
690dc03131 add chain_sum curriculum unit test 2025-02-10 22:09:18 +01:00
Zafir Stojanovski
178895ab1b Power Function (#102)
* power function dataset + tests
2025-02-10 22:04:58 +01:00
Zafir Stojanovski
696fdf8be7 Merge branch 'main' of https://github.com/open-thought/reasoning-gym into env/matrix-manipulation 2025-02-10 20:40:41 +01:00
Andreas Koepf
357a89fe8c Add attributes for curriculum
Co-authored-by: EduardDurech <39579228+EduardDurech@users.noreply.github.com>
2025-02-10 18:58:07 +01:00
Adefioye
767c34297f Add score_answer method to word_ladder (#93)
* Add score_answer method to word_ladder
* add unit test for WordLadderDataset::score_answer()

---------

Co-authored-by: Andreas Koepf <andreas.koepf@provisio.com>
2025-02-10 15:15:23 +01:00
Zafir Stojanovski
111f4c9170 matrix manipulation 2025-02-10 13:51:39 +01:00
Andreas Köpf
3150f9d9aa Merge pull request #97 from rishabhranawat/eval-v1
[eval-basic] initial scripts for evaluating models on reasoning gym
2025-02-10 11:59:49 +01:00
rishabhranawat
03f87dbc07 [eval-basic] remove large results files, add gitignore, only leave summary 2025-02-09 22:52:10 -08:00
rishabhranawat
2308ed99fb [eval-basic] run precommit formatting 2025-02-09 22:40:45 -08:00
rishabhranawat
94f07ed35d [eval-basic] initial scripts for evaluating models on reasoning gym 2025-02-09 22:36:27 -08:00
Oliver
a53073278a Remove rng param 2025-02-09 21:26:03 +00:00
Oliver
0627f2b02d Greatly speed up solver 2025-02-09 21:23:53 +00:00
Andreas Koepf
c59db00196 reduce default zero probability for binary matrix 2025-02-09 20:05:56 +01:00
Andreas Köpf
0605c0cbe4 Merge pull request #91 from zafstojano/env/binary-matrix
Binary Matrix
2025-02-09 19:55:36 +01:00
Andreas Köpf
7444f774c6 Merge pull request #92 from rishabhranawat/poly-reward
Add score_answer() for PolynomialEquationsDataset
2025-02-09 19:30:24 +01:00
rishabhranawat
04b3323844 [poly-reward] run pre-commit hooks 2025-02-09 07:30:18 -08:00
Zafir Stojanovski
e5862371ed update instruction and shuffle numbers 2025-02-09 13:00:46 +01:00
Andreas Köpf
70c731e9fb Merge pull request #94 from zafstojano/fix/prime-factorization-scoring
fix(env): Prime Factorization scoring
2025-02-09 12:02:11 +01:00
Zafir Stojanovski
6cc5d0dd63 normalize answer and partial reward 2025-02-09 11:13:23 +01:00
rishabhranawat
7a6f7ea9da [poly-reward] minor updates to the docstrings 2025-02-08 21:41:18 -08:00
rishabhranawat
0f4ab53bd3 Merge branch 'main' of https://github.com/rishabhranawat/reasoning-gym into poly-reward 2025-02-08 21:37:21 -08:00
rishabhranawat
0dd4c05897 [poly-reward] add a greedy strategy scoring function for polynomial equations 2025-02-08 21:36:21 -08:00
Zafir Stojanovski
89fd56f8e9 RotateMatrix typo 2025-02-09 01:11:06 +01:00
Zafir Stojanovski
f7836e17d0 binary matrix 2025-02-09 01:10:57 +01:00
Andreas Koepf
0c7fbb5001 bump version 2025-02-09 00:39:48 +01:00
Andreas Koepf
04bffd8f59 update GALLERY.md after merging knight_swap 2025-02-09 00:35:56 +01:00
Andreas Köpf
17ea7dd975 Merge pull request #89 from JeanKaddour/feat-swap-knights-puzzles
Feat swap knights puzzles
2025-02-09 00:33:48 +01:00
Andreas Köpf
f5a6dabb8b Merge pull request #90 from open-thought/arc_agi_1_dataset
ARC-AGI-1 dataset with augmentations
2025-02-09 00:19:20 +01:00
Andreas Koepf (aider)
ec8036c099 feat: Add configurable rotation and mirror augmentation variants 2025-02-09 00:16:41 +01:00
Andreas Koepf
b73040b066 refactor: Reorganize ArcAgiConfig class attributes for better readability 2025-02-09 00:12:59 +01:00
Andreas Koepf
e56316ebb2 formatting 2025-02-09 00:04:42 +01:00
Andreas Koepf (aider)
8d8d85e6b2 fix: Add missing Callable import to arc_agi.py 2025-02-08 23:59:53 +01:00
Andreas Koepf (aider)
cdb9d8d8f8 feat: Add configurable augmentations to ArcAgiDataset with consistent application 2025-02-08 23:59:45 +01:00
Andreas Koepf
1795cd815c add rotate, mirror & color-mapping augmentation functions 2025-02-08 23:51:38 +01:00
Andreas Koepf (aider)
f72bd8d6a5 test: Add comprehensive unit tests for ArcAgiDataset 2025-02-08 23:20:45 +01:00
Andreas Koepf
4e49806d22 add ArcAgiDataset class, fix score_entry() metadata params 2025-02-08 23:18:18 +01:00