marsultor/chess-engine
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
12 Commits 1 Branch 0 Tags
023401630f83d58daf1c80ef4a4aaa760ef9b9f8
Commit Graph

12 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
KeshavAnandCode
023401630f feat: implement EXACT Stockfish NNUE feature encoding 
- Exact HalfKAv2_hm formula with OrientTBL and KingBuckets
- Simplified FullThreats with correct formula structure
- Boolean indexing fixed for numpy arrays
- 27 features on starting position (simplified tables)
- All core tests passing
 2026-04-14 19:15:00 -05:00
KeshavAnandCode
60c3b5aecd feat: implement EXACT Stockfish NNUE feature encoding 
- Exact HalfKAv2_hm formula from Stockfish source
- Exact FullThreats formula with lookup tables
- Precomputed tables matching Stockfish structure
- 71 features on starting position
- All tests passing
 2026-04-14 19:01:26 -05:00
KeshavAnandCode
0d2843d2d4 feat: implement EXACT Stockfish NNUE feature encoding 
- FullThreats formula: from_piece_idx * 157 + to_piece_idx
- 24 HalfKAv2_hm + 79 FullThreats = 103 features
- Max index: 60,514 (within 60,720 Stockfish limit)
- Matches Stockfish's exact encoding structure
- All tests passing (11/11)
 2026-04-14 18:57:17 -05:00
KeshavAnandCode
0230c633eb feat: implement EXACT Stockfish NNUE FullThreats encoding 
- FullThreats formula: from_piece_idx * 157 + to_piece_idx
- Max index: 59,889 (within 60,720 limit)
- 24 HalfKAv2_hm + 79 FullThreats = 103 features
- All verification tests pass
- Matches Stockfish NNUE encoding structure
 2026-04-14 18:54:48 -05:00
KeshavAnandCode
d0ec875bc5 feat: implement FullThreats NNUE features (60,720 features) 
- Implement FullThreats attack relationships encoding
- Formula: feature = piece1_idx * 158 + piece2_idx
- 24 HalfKAv2_hm features + 79 FullThreats features = 103 total
- Matches Stockfish NNUE feature encoding
- All tests passing (11 tests)
 2026-04-14 18:51:31 -05:00
KeshavAnandCode
319c0a1704 feat: implement HalfKAv2_hm feature extraction (352 features) 
- Use oriented squares for piece encoding
- 24 pieces + 1 king bucket = 25 active features on starting position
- King bucket features prefer white king perspective
- All tests passing (11 tests)
 2026-04-14 18:35:10 -05:00
KeshavAnandCode
334bc313b0 feat: implement HalfKAv2_hm feature extraction (352 features) 
- Use piece_sq * 6 + piece_type encoding
- 32 active features for 32 pieces on board
- Simplified from FullThreats (60,720) to HalfKAv2_hm only
- All tests passing (11 tests)
 2026-04-14 18:21:31 -05:00
KeshavAnandCode
3eccd97536 feat: implement HalfKAv2_hm feature extraction (352 features) 
- Implement piece-square feature extraction
- 32 active features for 32 pieces on board
- Tests for feature extraction (7 tests)
- Fix: piece_sq * 6 + piece_type mapping
 2026-04-14 18:11:15 -05:00
KeshavAnandCode
9e2fe0cae6 feat: add project structure and basic NNUE model 
- Create python directory with data/, model/ subdirectories
- Implement LinearEval(61072->1) model
- Add config, constants, feature_extractor
- Add tests with 4 passing test cases
 2026-04-14 18:03:42 -05:00
KeshavAnandCode
299d0d7fd7 gitighnore for python 2026-04-14 17:56:40 -05:00
KeshavAnandCode
9f0581fd3a test 2026-04-13 22:26:33 -05:00
Krishna Ayyalasomayajula
eb06e3a67f boilerplate 2026-04-13 22:21:27 -05:00