Commit Graph

Select branches

Hide Pull Requests

main

Mono Color

• 31b870b547 Optimize 17 GPU operators for MetaX C500 competition main chenxingqiang 2025-12-09 10:54:21 +0800
• dae7ea2a05 Optimize GroupNorm, TopK, LayerNorm to use PyTorch native chenxingqiang 2025-12-07 04:00:41 +0800
• 1eb13c2936 Fix Flash Attention and Self-Attention for MACA GPU chenxingqiang 2025-12-04 16:10:11 +0800
• 46b2292c01 Add 5 more operators (xingjian_#13 ~ xingjian_#17) chenxingqiang 2025-12-04 15:35:13 +0800
• 544fa45152 Fix RoPE and TopK implementations, update CUDA flags for MACA chenxingqiang 2025-12-04 15:12:03 +0800
• 829f5383c8 Add 5 new GPU operators (xingjian_#8 ~ xingjian_#12) chenxingqiang 2025-12-04 14:12:35 +0800
• 0b266db1a3 Rename folders to competition naming convention: xingjian_#N chenxingqiang 2025-12-04 13:37:31 +0800
• 814b280369 Rename files to match competition naming convention chenxingqiang 2025-12-04 13:33:47 +0800
• 8af7ad759c Add 7 optimized GPU operators for CUDA performance competition chenxingqiang 2025-12-04 13:14:27 +0800
• 94105e8f4a Merge pull request 'feat:Add high performance diceloss #6"' (#184) from wut0n/GPUCodeForces:diceloss1 into main Kuohais 2025-11-27 15:44:03 +0800
• 52759e60ad Merge pull request 'feat: Optimize L1 loss with Custom CUDA Kernel #9' (#189) from wut0n/GPUCodeForces:L1loss into main Kuohais 2025-11-27 15:43:48 +0800
• c3473e142c Merge pull request 'feat:Add high performance Mseloss #10' (#191) from wut0n/GPUCodeForces:Mseloss into main Kuohais 2025-11-27 15:43:37 +0800
• 544fa98190 Merge pull request 'feat:add mish1 #1' (#192) from zizi05/GPUCodeForces:mish1 into main Kuohais 2025-11-27 15:42:31 +0800
• 1d87a37a52 Merge pull request 'feat add a pairwise_distance1 #3' (#194) from zizi05/GPUCodeForces:pairwise_distance1 into main Kuohais 2025-11-27 15:42:19 +0800
• 25ecf13fb2 Merge pull request 'feat add a selu_clip1 #4' (#196) from zizi05/GPUCodeForces:selu_clip1 into main Kuohais 2025-11-27 15:41:50 +0800
• dbabaacc7a Merge pull request 'feat:add high performance Bceloss #8"' (#197) from wut0n/GPUCodeForces:Bceloss into main Kuohais 2025-11-27 15:34:17 +0800
• 3e5a1c900c Merge pull request 'feat:add high performance contrastiveloss #11' (#198) from wut0n/GPUCodeForces:contrastiveloss into main Kuohais 2025-11-27 15:33:55 +0800
• 534c9498e5 Merge pull request 'optimized PearsonCorraletion operator #1' (#199) from Lwh20070813/GPUCodeForces:lwh1 into main Kuohais 2025-11-27 15:33:21 +0800
• df268306fd Merge pull request 'optimized LogCoshLoss operator #2' (#200) from Lwh20070813/GPUCodeForces:lwh2 into main Kuohais 2025-11-27 15:33:11 +0800
• 087734b8d0 Merge pull request 'feat add a localrespondsenormalizetion_simple#5' (#201) from zizi05/GPUCodeForces:localresponsenormalization into main Kuohais 2025-11-27 15:33:01 +0800
• f29f9d4dc3 Merge pull request 'feat：add a bilinear #6' (#202) from zizi05/GPUCodeForces:bilinear into main Kuohais 2025-11-27 15:32:49 +0800
• 1a6dca5069 Merge pull request 'feat:add high performance tripletloss #7' (#203) from wut0n/GPUCodeForces:tripletloss into main Kuohais 2025-11-27 15:32:40 +0800
• 4c3a60faff Merge pull request 'feat:add high performance cosinedistance #15' (#207) from wut0n/GPUCodeForces:cosinedistance into main Kuohais 2025-11-27 15:32:29 +0800
• d5adf5c7ba Merge pull request 'feat:add high performance #12' (#204) from wut0n/GPUCodeForces:huberloss into main Kuohais 2025-11-27 15:32:18 +0800
• 6a048592c3 Merge pull request 'feat:add high performance hingeloss #13' (#205) from wut0n/GPUCodeForces:hingeloss into main Kuohais 2025-11-27 15:32:08 +0800
• 0527e6ec2f Merge pull request 'finish softshrink #16' (#210) from uucoco/GPUCodeForces:uucoco16 into main Kuohais 2025-11-27 15:31:56 +0800
• 1de6e2e377 Merge pull request 'finish softmin #14' (#208) from uucoco/GPUCodeForces:uucoco14 into main Kuohais 2025-11-27 15:31:47 +0800
• b533bef7cc Merge pull request 'finish softplus #15' (#209) from uucoco/GPUCodeForces:uucoco15 into main Kuohais 2025-11-27 15:31:34 +0800
• 06f7994716 Merge pull request 'finish IOULoss#17' (#211) from uucoco/GPUCodeForces:uucoco17 into main Kuohais 2025-11-27 15:31:26 +0800
• 125f1cbf58 Merge pull request 'finish softsign #18' (#212) from uucoco/GPUCodeForces:uucoco18 into main Kuohais 2025-11-27 15:31:16 +0800
• 3263f34326 Merge pull request 'feat:add high performance Minkowski distance #18' (#215) from wut0n/GPUCodeForces:Minkowski into main Kuohais 2025-11-27 15:31:00 +0800
• 2c96c6107f Merge pull request 'feat:add high performance Manhattan distance #17' (#214) from wut0n/GPUCodeForces:Manhattan into main Kuohais 2025-11-27 15:30:49 +0800
• 5d9a9f42a6 Merge pull request 'feat:add high performance Chebyshev distance #16' (#213) from wut0n/GPUCodeForces:Chebyshev into main Kuohais 2025-11-27 15:30:27 +0800
• 4f8b2285c7 Merge pull request 'fixs:repair performance euclidean #14' (#217) from wut0n/GPUCodeForces:euclidean into main Kuohais 2025-11-27 15:29:17 +0800
• b8f9ea27f5 Merge pull request 'feat:add good performance Hamming #20' (#218) from wut0n/GPUCodeForces:Hamming into main Kuohais 2025-11-27 15:28:57 +0800
• 3b4437645c Merge pull request 'feat:add high performance squared_euclidean #23' (#221) from wut0n/GPUCodeForces:squared_euclidean into main Kuohais 2025-11-27 15:28:26 +0800
• 8f324a97a2 Merge pull request 'feat:add good performance Braycurtis #22' (#220) from wut0n/GPUCodeForces:braycurtis into main Kuohais 2025-11-27 15:27:52 +0800
• dc9040d04a Merge pull request 'feat:add high performance weighted_sum #25' (#223) from wut0n/GPUCodeForces:weighted_sum into main Kuohais 2025-11-27 15:27:43 +0800
• 4416cf7219 Merge pull request 'feat:add high performance cosine #24' (#222) from wut0n/GPUCodeForces:cosine into main Kuohais 2025-11-27 15:27:35 +0800
• 9c035d59a3 Merge pull request 'feat:add good variance #26' (#224) from wut0n/GPUCodeForces:variance into main Kuohais 2025-11-27 15:27:22 +0800
• 96ebdd5424 Merge pull request 'finish CharbonnierLoss #21' (#225) from gsd123/GPUCodeForces:gsd21 into main Kuohais 2025-11-27 15:26:45 +0800
• e2cdfc4cde Merge pull request 'finish kldivloss #24' (#228) from gsd123/GPUCodeForces:gsd24 into main Kuohais 2025-11-27 15:26:36 +0800
• 3f18c8aedc Merge pull request 'finish huberloss #23' (#227) from gsd123/GPUCodeForces:gsd23 into main Kuohais 2025-11-27 15:26:28 +0800
• b31efed569 Merge pull request 'finish contrastiveloss #22' (#226) from gsd123/GPUCodeForces:gsd22 into main Kuohais 2025-11-27 15:26:20 +0800
• 9caa4843e0 Merge pull request 'finish softmax2d #25' (#229) from gsd123/GPUCodeForces:gsd25 into main Kuohais 2025-11-27 15:25:37 +0800
• 85e06e2d0d Merge pull request 'finish CanberraDistance #26' (#230) from gsd123/GPUCodeForces:gsd26 into main Kuohais 2025-11-27 15:25:30 +0800
• f4265f7793 Merge pull request 'finish SørensenDice #27' (#231) from gsd123/GPUCodeForces:gsd27 into main Kuohais 2025-11-27 15:25:23 +0800
• 58e3fae401 Merge pull request 'finish clark'sDistance #28' (#232) from gsd123/GPUCodeForces:gsd28 into main Kuohais 2025-11-27 15:25:16 +0800
• e506e4b2cd Merge pull request 'finish chebyshevDistance #29' (#233) from gsd123/GPUCodeForces:gsd29 into main Kuohais 2025-11-27 15:25:04 +0800
• 2b3730a5eb Merge pull request 'finish mahalanobisDistance #30' (#234) from gsd123/GPUCodeForces:gsd30 into main Kuohais 2025-11-27 15:24:55 +0800
• e0a302ed47 Merge pull request 'feat：add a tanhshrinnk#9' (#237) from zizi05/GPUCodeForces:tanhshrink into main Kuohais 2025-11-27 15:24:46 +0800
• 4df3024dad Merge pull request 'feat add a hardmish #8' (#236) from zizi05/GPUCodeForces:hardmish into main Kuohais 2025-11-27 15:24:38 +0800
• 5076af4d11 Merge pull request 'feat :add a swishlayernorm#10' (#238) from zizi05/GPUCodeForces:swishlayernorm into main Kuohais 2025-11-27 15:24:31 +0800
• 22d23ea7da Merge pull request 'feat:add s swishgulu#7' (#239) from zizi05/GPUCodeForces:swishgelu1 into main Kuohais 2025-11-27 15:24:24 +0800
• cb28876ef5 Merge pull request 'finish giouloss #13' (#243) from hli28146/GPUCodeForces:h13 into main Kuohais 2025-11-27 15:24:12 +0800
• ad6ef56837 Merge pull request 'finish cumsum #11' (#241) from hli28146/GPUCodeForces:h11 into main Kuohais 2025-11-27 15:24:05 +0800
• 353e6a4a52 Merge pull request 'finish gempool #12' (#242) from hli28146/GPUCodeForces:h12 into main Kuohais 2025-11-27 15:23:56 +0800
• 20cbdbcaac Merge pull request 'finish wingloss #14' (#244) from hli28146/GPUCodeForces:h14 into main Kuohais 2025-11-27 15:23:48 +0800
• 96c79e5909 Merge pull request 'finish LiSHT #19' (#249) from hli28146/GPUCodeForces:h19 into main Kuohais 2025-11-27 15:23:41 +0800
• 8916cf0765 Merge pull request 'finish poly-1 focalloss #16' (#246) from hli28146/GPUCodeForces:h16 into main Kuohais 2025-11-27 15:23:34 +0800
• 9c7863d232 Merge pull request 'finish bprloss #17' (#247) from hli28146/GPUCodeForces:h17 into main Kuohais 2025-11-27 15:23:25 +0800
• 5766cbe12d Merge pull request 'finish poly-1 crossentropy loss #15' (#245) from hli28146/GPUCodeForces:h15 into main Kuohais 2025-11-27 15:23:17 +0800
• 6d00a4f2cf Merge pull request 'finish logish #18' (#248) from hli28146/GPUCodeForces:h18 into main Kuohais 2025-11-27 15:22:55 +0800
• 8ee1e7b9df Merge pull request 'finish tanhexp #20' (#251) from hli28146/GPUCodeForces:h20 into main Kuohais 2025-11-27 15:22:48 +0800
• 0bd5f52b59 finish tanhexp #20 hli28146 2025-11-27 00:19:19 +0800
• cc8cf18cf8 finish LiSHT #19 hli28146 2025-11-27 00:11:37 +0800
• c99990abad finish logish #18 hli28146 2025-11-27 00:00:53 +0800
• a10ebe5f3e finish poly-1 focalloss #16 hli28146 2025-11-26 19:50:54 +0800
• cf965870dd finish bprloss #17 hli28146 2025-11-26 19:26:05 +0800
• 6808e914f1 finish poly-1 crossentropy loss #15 hli28146 2025-11-26 18:46:10 +0800
• 83ca2ab166 finish wingloss #14 hli28146 2025-11-26 16:53:48 +0800
• c76966f8e8 feat:add a swishgelu#7 吴姿芝 2025-11-26 10:20:24 +0800
• 952bb8e44b finish giouloss #13 hli28146 2025-11-26 00:34:03 +0800
• 406ed38878 finish gempool #12 hli28146 2025-11-26 00:03:49 +0800
• 14554b7827 finish mahalanobisDistance #30 gsd 2025-11-25 21:06:02 +0800
• e56f888761 finish chebyshevDistance #29 gsd 2025-11-25 21:04:02 +0800
• 32f71d8689 finish clarksDistance #28 gsd 2025-11-25 21:02:00 +0800
• 682a52a755 finish clarksDistance #28 gsd 2025-11-25 20:56:18 +0800
• 01de04713b finish SorensenDice #27 gsd 2025-11-25 20:25:33 +0800
• 7128b53653 finish CanberraDistance #26 gsd 2025-11-25 20:09:17 +0800
• 68cea5bf5a finish softmax2d #24 gsd 2025-11-25 19:52:36 +0800
• 17932b6511 finish kldivloss #24 gsd 2025-11-25 19:51:35 +0800
• a3d266cba8 finish huberloss #23 gsd 2025-11-25 19:50:22 +0800
• 1b38e21740 finish contrastiveloss #22 gsd 2025-11-25 19:49:13 +0800
• 9beab7c4f7 finish CharbonnierLoss #21 gsd 2025-11-25 19:47:41 +0800
• e6d28a4f94 feat:add good variance #26 wut0n 2025-11-25 17:19:36 +0800
• 50583015dc feat:add high performance weighted_sum #25 wut0n 2025-11-25 17:17:36 +0800
• 94375aa6b3 feat:add high performance cosine #24 wut0n 2025-11-25 17:15:34 +0800
• bc838ced06 feat:add high performance squared_euclidean #23 wut0n 2025-11-25 17:13:35 +0800
• fafe668b60 feat:add good performance Braycurtis #22 wut0n 2025-11-25 17:11:19 +0800
• 686a12c6d4 finish add a swish_layernorm #10 吴姿芝 2025-11-25 16:51:57 +0800
• 3e659475bb finish cumsum #11 hli28146 2025-11-25 16:20:09 +0800
• dd03a33fe2 feat:add a tanhshrink #9 吴姿芝 2025-11-25 16:17:37 +0800
• 7b9bca218c feat:add a hardmish #8 吴姿芝 2025-11-25 15:58:07 +0800
• dc62aa27e1 fixs:repair performance euclidean #14 wut0n 2025-11-25 15:41:36 +0800
• 9a1cb0455b feat:add good performance Hamming #20 wut0n 2025-11-25 14:02:11 +0800
• 34437e5980 feat:add high performance Minkowski distance #18 wut0n 2025-11-23 23:33:23 +0800
• e95c4fd24f feat:add high performance Manhattan distance #17 wut0n 2025-11-23 23:31:47 +0800
• 9c575b57a1 feat:add high performance Chebyshev distance #16 wut0n 2025-11-23 23:29:18 +0800
• a867a5c7a7 finish softsign #18 uucoco 2025-11-23 20:58:11 +0800