Brutex
/
cavis
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Commit Graph

  • ca881a987a
         non-inplace pydatavec transform processes (#8326) Max Pumperla 2019-10-30 15:59:54 +0100
  • 95f7ad7b94
         Shugeo suppression overlaps (#9) shugeo 2019-10-30 13:43:45 +0200
  • cea72582b8
         pydatavec examples link was recognized as relative (#8321) Max Pumperla 2019-10-30 09:03:27 +0100
  • 5a4d2e8b31
         [WIP] SVD (#16) raver119 2019-10-28 12:31:01 +0300
  • 029a69a835
         Shyrma bn mkl bp (#14) Yurii Shyrma 2019-10-26 14:14:21 +0300
  • d333d29099
         SameDiff cleanup and fixes (#12) Alex Black 2019-10-26 12:38:08 +1100
  • d98784197a
         [WIP] Some optimizations in dl4j (#11) Alexander Stoyakin 2019-10-24 14:15:50 +0300
  • 730442ae21
         Remove bad import (#13) Alex Black 2019-10-24 20:51:15 +1100
  • 3f0b4a2d4c
         SameDiff execution, TF and memory management overhaul (#10) Alex Black 2019-10-23 21:19:50 +1100
  • f31661e13b
         Merge pull request #7 from KonduitAI/asto_nd4s_10172019 Alexander Stoyakin 2019-10-23 12:11:25 +0300
  • 35e6ffede4
         Merge pull request #6 from KonduitAI/shyrma_broadcast raver119 2019-10-22 07:56:04 +0300
  • 8f3eaebda5 - replace condition isScalar() by condition length ==1 in some NDArray methodds Yurii 2019-10-21 16:25:13 +0300
  • 775402c934
         Merge pull request #5 from KonduitAI/asto_nd4s_build_10192019 Alexander Stoyakin 2019-10-19 21:57:02 +0300
  • b556519e15 Fix compilation. Alexander Stoyakin 2019-10-19 21:48:43 +0300
  • 7f06df5826
         Small URL fix (#8297) Alex Black 2019-10-18 11:25:58 +1100
  • a4984d52d7
         Merge pull request #4 from KonduitAI/shyrma_lstm raver119 2019-10-17 20:58:06 +0300
  • 99be467f76 - minor change in recurrent.h Yurii 2019-10-17 20:46:51 +0300
  • 70bd925abd - write 2 versions of new lstmLayer: one is based on own code, second uses mkl dnn api Yurii 2019-10-17 20:44:52 +0300
  • 630bb3c9b6
         Merge pull request #2 from KonduitAI/asto_ops_wrapper Alexander Stoyakin 2019-10-16 20:21:50 +0300
  • 9e5799847a TF names fixed. Alexander Stoyakin 2019-10-16 19:50:18 +0300
  • 502bedf5d5 Register ops for TF import. Alexander Stoyakin 2019-10-16 19:39:04 +0300
  • ec722b20ee TF names added Alexander Stoyakin 2019-10-16 19:29:19 +0300
  • 99d77e1384 Ops exported for sameDiff Alexander Stoyakin 2019-10-16 19:16:47 +0300
  • 3662657d5c
         Merge pull request #1 from KonduitAI/shugeo_gamma shugeo 2019-10-16 18:49:33 +0300
  • 24a2b2933f Added gamma and lgamma functions. shugeo 2019-10-16 18:22:18 +0300
  • 96a9a1a733 Fixed output from operation. Alexander Stoyakin 2019-10-16 18:07:52 +0300
  • 7617682a46 Added declarations for igamma and igammac ops. shugeo 2019-10-16 14:45:10 +0300
  • 478a0c1f97 Added igamma and igammac broadcastable ops implementations and tests. shugeo 2019-10-16 14:02:53 +0300
  • 7103aca8c5 Added broadcastable IGamma and IGammac ops. shugeo 2019-10-16 13:58:32 +0300
  • f90e6da97e Added nd4j_gamma, nd4j_igamma and nd4j_igammac functions. shugeo 2019-10-16 13:53:31 +0300
  • c4307384f3 Fixed shape for muli Alexander Stoyakin 2019-10-16 12:59:25 +0300
  • d5002b14c7 New ops wrappers Alexander Stoyakin 2019-10-16 12:59:08 +0300
  • b9a4b0f25f
         Update links (#8292) Alex Black 2019-10-16 12:59:52 +1100
  • 171ce51f46 RL4J: Use Nd4j Random instead of java.util.Random (#8282) Alexandre Boulanger 2019-10-15 21:56:24 -0400
  • df2448613e Added gamma distribution functions. shugeo 2019-10-15 20:00:07 +0300
  • 2d750b69e5 Merge remote-tracking branch 'konduit/master' AlexDBlack 2019-10-14 17:21:23 +1100
  • d5568df7b3
         Merge pull request #10 from KonduitAI/shugeo_fake_quant2 raver119 2019-10-10 18:40:02 +0300
  • ace65355c5 Added doc for fake_quant_with_min_max* op helpers cuda implementations. shugeo 2019-10-10 18:35:28 +0300
  • c890de5a7b Added doc for fake_quant_with_min_max* op helpers implementations. shugeo 2019-10-10 18:31:17 +0300
  • c3f755d975 Refactored helpers both for cuda and cpu platforms. shugeo 2019-10-10 18:02:49 +0300
  • a09cb5e2be Added doc for fake_quant_with_min_max_per_channel op declaration. shugeo 2019-10-10 17:13:33 +0300
  • 92636b0b86 Eliminated waste operator. shugeo 2019-10-10 17:08:59 +0300
  • d5b352273d Implementation of cuda kernel for fake_quant_with_min_max_vars_per_channels op. Final revision. shugeo 2019-10-10 16:51:29 +0300
  • 02d8616692 Implementation of cuda kernel for fake_quant_with_min_max_vars_per_channels op. shugeo 2019-10-10 16:40:56 +0300
  • 3504b0cda9 Implemented fake_quant_with_min_max_vars_per_channel fop cuda helper. The first working revision. shugeo 2019-10-10 15:44:50 +0300
  • 753565145c Refactored fake_quant_with_min_max_vars op cuda implementation. shugeo 2019-10-10 14:00:49 +0300
  • c13e945a96 Fixed fake_quant_with_min_max_vars op and tests. shugeo 2019-10-10 13:23:11 +0300
  • fa8105bf0f
         Merge pull request #9 from KonduitAI/asto_nd4s_indexing Alexander Stoyakin 2019-10-10 11:03:06 +0300
  • 156fba4f77 Minor fix Alexander Stoyakin 2019-10-10 10:49:07 +0300
  • be70bea359 Operator :: for INDArray Alexander Stoyakin 2019-10-10 10:43:32 +0300
  • 3c0c59ab88 Refactored fake_quant_with_min_max_vars op. shugeo 2019-10-09 22:09:33 +0300
  • 352f1eee80 Implemented fake_quant_with_min_max_per_channel helper for cpu platform. The first approach. shugeo 2019-10-09 21:39:59 +0300
  • 0de8523ebd Indexing syntax changed Alexander Stoyakin 2019-10-09 16:07:20 +0300
  • d0cbd33b0e Added input checks for op. shugeo 2019-10-09 15:52:13 +0300
  • 3a89e51811 Added tests for fake_quant_with_min_max_vars_per_channel op. shugeo 2019-10-09 13:38:18 +0300
  • f515151f5f
         Merge pull request #8 from KonduitAI/asto_nd4s Alexander Stoyakin 2019-10-09 11:39:51 +0300
  • 3aa51e210a RL4J: Extract TD Target calculations (StandardDQN and DoubleDQN) (#8267) Alexandre Boulanger 2019-10-08 20:14:47 -0400
  • 8472782d7f More tests Alexander Stoyakin 2019-10-08 19:08:43 +0300
  • cb56b0b06a The first approach for fake_quant_with_min_max_vars_per_channel op implementation. shugeo 2019-10-08 19:00:41 +0300
  • 4956048a11 Tests added Alexander Stoyakin 2019-10-08 17:38:43 +0300
  • b36b3fa1aa Some sugar Alexander Stoyakin 2019-10-08 16:41:12 +0300
  • dbabd11f83
         Merge pull request #7 from KonduitAI/shugeo_image_bordering shugeo 2019-10-08 15:47:37 +0300
  • 8fe5a1fa96 The working implementation of draw_bounding_boxes op. shugeo 2019-10-08 15:42:27 +0300
  • 50b13fadc8 nd4j-api cleanup. (#8273) Robert Altena 2019-10-08 19:48:22 +0900
  • 30a8af566c The first working implementation of cuda kernel for draw_bounding_boxes op helper. shugeo 2019-10-08 13:45:18 +0300
  • ae09cfee32 Next approach of cuda imlementation for draw_bounding_boxes op helper. shugeo 2019-10-08 00:09:46 +0300
  • 6cf3a8fa9c Refactored cpu implementatio and added cuda aproach. shugeo 2019-10-07 17:51:07 +0300
  • 7a1e327516 Indexing demo Alexander Stoyakin 2019-10-07 16:59:47 +0300
  • 78443ffebf Working implementation of draw_bounding_boxes op for cpu. shugeo 2019-10-07 15:04:44 +0300
  • 16a66a65e3 Added helper declaration for draw_bounding_boxes op. shugeo 2019-10-04 21:16:34 +0300
  • 53a2ebddbe Added test and helpers for draw_bounding_boxes op both cpu and cuda related. shugeo 2019-10-04 20:46:26 +0300
  • 8f70b4441f draw_bounding_boxes op implementation. Inital revision. shugeo 2019-10-04 18:32:21 +0300
  • b8f2a83a5a
         Merge pull request #5 from KonduitAI/shugeo_divnonan_full raver119 2019-10-04 10:33:04 +0300
  • 908e4c4912 Added implementation for divide_no_nan op and tests. shugeo 2019-10-04 10:29:15 +0300
  • 2381c51480
         Merge pull request #4 from KonduitAI/revert-3-shugeo_div_no_nan raver119 2019-10-03 20:26:15 +0300
  • cff26f13c5
         Revert "Implement divide_no_nan op." raver119 2019-10-03 20:25:52 +0300
  • bbfa41cbc8
         Merge pull request #3 from KonduitAI/shugeo_div_no_nan raver119 2019-10-03 18:27:12 +0300
  • 6eaca179d6 Implement divide_no_nan op. shugeo 2019-10-03 18:22:17 +0300
  • d2e98564d4
         Merge pull request #2 from KonduitAI/shugeo_bincast raver119 2019-10-03 14:27:24 +0300
  • 1f4ad08305 refactoring activations. (#8261) Robert Altena 2019-10-03 19:35:27 +0900
  • 130ee25682 Implemented compare_and_bitpack op. shugeo 2019-10-03 10:57:48 +0300
  • 75ad3c8153 Fixed test names. shugeo 2019-10-02 19:05:26 +0300
  • f3e42173ef Refactored buffer copying to avoid wrong usage of buffers. shugeo 2019-10-02 16:51:09 +0300
  • 1c6173d218 Added implementation of bitcast op. shugeo 2019-10-02 15:04:59 +0300
  • a27e61553a Added tests and fixed op name. shugeo 2019-10-02 15:04:28 +0300
  • 863ff76878 Added declaration for bincast op. shugeo 2019-10-02 12:17:00 +0300
  • 62326542e9
         Merge pull request #1 from KonduitAI/shugeo_adjust raver119 2019-10-01 14:15:54 +0300
  • afeb524238 Refactored implementation for adjust_contrast ops. shugeo 2019-10-01 14:13:09 +0300
  • 1575c704ae Added implementation for adjust_contrast_v2 op and tests. shugeo 2019-10-01 11:44:27 +0300
  • 44a8d19ac6
         [WIP] Broadcast changes (#8257) raver119 2019-10-01 09:10:19 +0300
  • 5959ff4795 RL4J: Fix QLearningDiscrete.setTarget() and add CartpoleNative (#8250) Alexandre Boulanger 2019-09-30 20:27:51 -0400
  • e06dfb5dcc Implementation of adjust_contrast op. shugeo 2019-09-30 18:24:12 +0300
  • d5e98afcef RL4J: Add VideoRecorder (#8106) Alexandre Boulanger 2019-09-30 00:40:32 -0400
  • f98f8be7b6
         SameDiff ops (#8247) Alex Black 2019-09-19 11:52:20 +1000
  • 59f1cbf0c6 RL4J - AsyncTrainingListener (#8072) Alexandre Boulanger 2019-09-18 21:28:13 -0400
  • d58a4b45b1 Fix Nadam updater clone missing schedule (#8243) Steven Lang 2019-09-17 21:56:49 +0200
  • 83d958d536 Sparse matrix refactoring. (#8238) Robert Altena 2019-09-18 04:56:29 +0900
  • 979ef13c0b fix build issues raver119 2019-09-13 11:55:13 +0300
  • 6b557f2441 Merge remote-tracking branch 'fork/master' AlexDBlack 2019-09-13 15:19:08 +1000
  • c1bcbdda18
         Build fix - tsne random method signature (#259) Alex Black 2019-09-13 13:40:46 +1000