https://github.com/Microsoft/CNTK

sort by:
Revision Author Date Message Commit Date
f408730 Merge branch 'master' into kedeng/mkldnn 01 May 2018, 16:59:55 UTC
bbf20c0 Restore number mistaken for version. 30 April 2018, 18:18:07 UTC
51ed19a Integrate db211f1d84965fd56f450cd4592f0c99b3b1d99c into master 27 April 2018, 01:46:53 UTC
de5621c Fixing Convolution layer doc to include dilation and groups arguments. 25 April 2018, 22:30:51 UTC
adf4826 Merge pull request #3156 from MalKeshar/patch-1 Allow to add extra cxxflags via env 25 April 2018, 17:59:22 UTC
98dc8e4 Integrate gloch/fix_native_proposallayerlib into master 24 April 2018, 21:23:28 UTC
fa8616a Merge branch 'master' into kedeng/mkldnn 24 April 2018, 21:07:09 UTC
d5db843 Add Linux OOBE support 24 April 2018, 20:35:46 UTC
575e21b Fix VS build when source code and VS are in different drives 24 April 2018, 20:09:24 UTC
d1e2c91 Fix native ProposalLayerLib dimensions indexes 24 April 2018, 14:16:20 UTC
e785a2b Allow to add some extra cxxflags via env Allow to add some extra cxxflags via system env. Use cases: 1) https://github.com/Microsoft/CNTK/issues/3155 export CXXFLAGS="-Wno-unused-variable -Wno-sign-compare" configure && make 2) Our libc/libstdc has been built with gcc-4.9. So we fail to build cntk in current state with gcc-5 in compatibility mode. https://gcc.gnu.org/onlinedocs/libstdc++/manual/using_dual_abi.html After patching cntk we have successful build: COMMON_FLAGS="-D_GLIBCXX_USE_CXX11_ABI=0" CXXFLAGS="-D_GLIBCXX_USE_CXX11_ABI=0 -Wno-unused-variable -Wno-sign-compare" configure && make (nvcc don't recongnise -W flags) 24 April 2018, 13:36:56 UTC
51b9c0e Fix conv2d bug when having conv2d over the static axes of sequences. 23 April 2018, 22:55:46 UTC
438f619 Merge pull request #3140 from Joe8Bit/master (chore) Cleanup ./README.md 23 April 2018, 17:35:44 UTC
cd03a71 Update README.md 23 April 2018, 17:35:27 UTC
086073e adding publication date 23 April 2018, 17:31:29 UTC
5a0f072 adding related works reference to DSSM tutorial 23 April 2018, 17:29:47 UTC
116b662 Add Windows OOBE support 23 April 2018, 17:11:52 UTC
ab0ef38 Update dependencies from VS2015 to VS2017 and Anaconda3 4.1.1 to 4.3.1 23 April 2018, 17:11:52 UTC
1964906 Fix PythonBindings build on machines with VS2017 23 April 2018, 17:11:52 UTC
69a95f8 Integrate liqun/RNNBlk2 into master 23 April 2018, 17:01:48 UTC
91a4727 Really fix the slice matrix issue 21 April 2018, 00:46:13 UTC
dda6e1d Fix ReLU engine for RNN as well 20 April 2018, 22:11:45 UTC
44f220f Disable plusEngine in RNN 20 April 2018, 22:05:01 UTC
3b0f731 Fixes 20 April 2018, 21:32:52 UTC
9cad712 undo to match master 20 April 2018, 21:00:08 UTC
2d4d116 Fix PlusNode 20 April 2018, 20:42:31 UTC
d0e8f4c Undo some more CopyToArray changes 20 April 2018, 20:00:06 UTC
3de2b85 Undo some formatting changes to better align with master 20 April 2018, 05:37:48 UTC
6170a01 Fix MathTests failure 20 April 2018, 00:12:29 UTC
401e511 Use TLS 1.2 in devInstall script. 20 April 2018, 00:09:26 UTC
373ad5c More merge fix 19 April 2018, 23:53:45 UTC
383a5bc Merge branch 'master' into kedeng/mkldnn 19 April 2018, 22:52:06 UTC
65c4738 Fix merge 19 April 2018, 22:47:53 UTC
90d1740 WIP fix after merge 19 April 2018, 22:02:17 UTC
e9cc87c Merge branch 'master' into kedeng/mkldnn_merge 19 April 2018, 21:38:49 UTC
baf3c66 change cprintf to use vsnprintf (#3124) change cprintf to use vsnprintf 19 April 2018, 20:05:55 UTC
0b78d90 (chore) Cleanup ./README.md * Reorganise the sections to better reflect the the main use cases of the `READMD.md` e.g. finding out about the project in general, rather than a changelog * Add more and cleanup markdown format (e.g. consistent single backticks for inline code, and python syntax highlighting for code blocks) * Gramatical/syntax fixes 19 April 2018, 16:46:11 UTC
133d346 Update documentation for 2.5.1 17 April 2018, 23:10:34 UTC
e237559 Bump release version to 2.5.1 17 April 2018, 23:06:45 UTC
78a5ff3 Add support to build wheel packages with third party libs 17 April 2018, 23:02:21 UTC
21f7869 Add mkldnn to JAVA dependencies 17 April 2018, 22:53:26 UTC
e3721e6 Remove hard tabs 17 April 2018, 21:29:18 UTC
0b9b253 Added group convolution to current_iteration .md 17 April 2018, 20:34:14 UTC
35e1f05 Fix build after merge 17 April 2018, 18:12:55 UTC
cf95e87 Merge pull request #3121 from zhenlinluo/kedeng/mkldnn fix math test error 17 April 2018, 18:10:47 UTC
900bfd0 Merge branch 'kedeng/mkldnn' into kedeng/mkldnn 17 April 2018, 18:08:14 UTC
532a173 Fix build after merge 17 April 2018, 06:00:07 UTC
db211f1 Remove unnecessary sys/io.h Machines with aarch64 architecure such as NVIDIA Jetson TX1 does not have sys/io.h. Without this commit, building libCntk.Core-2.5.so fails on a aarch64 machine. It should be safe to completely remove the inclusion of sys/io.h, since it seems that functions in the header are not used by CNTK. 16 April 2018, 01:16:59 UTC
1ffdc6e Integrate sptiwari/shufflenet into master 14 April 2018, 05:53:15 UTC
9375ae5 fix math test error Signed-off-by: Luo <zhenlin.luo@intel.com> 14 April 2018, 01:43:58 UTC
cf839dc Fixing ONNX BatchNormalization implementation to match spec. 13 April 2018, 23:14:34 UTC
2e7bbaa Fix windows script to install cntk from binary drop 13 April 2018, 22:11:48 UTC
cd9285c make GRU/RNNStep blocks, issue 3084 to expose BlockArgumentsMapping 13 April 2018, 21:12:45 UTC
89db005 Integrate kedeng/RNNfp16_2 into master 12 April 2018, 20:23:41 UTC
bf97391 Merge pull request #3117 from zhenlinluo/kedeng/mkldnn Kedeng/mkldnn 12 April 2018, 03:20:32 UTC
ac1c15e add bn-relu fuse into Resnet bs for mkldnn benchmark Signed-off-by: ZhenlinLuo <zhenlin.luo@intel.com> 11 April 2018, 22:49:37 UTC
5e0856e Fix bugs for fp16 in RNN and BMUF Note that sparse embedding is still not working yet. 11 April 2018, 22:35:08 UTC
8bd55b3 Donot abort early if some ranks failed to initial NCCL 11 April 2018, 22:06:56 UTC
09e25a4 Moving group convolution implementation to use cuDNN7 and MKL2017 APIs. 11 April 2018, 17:26:15 UTC
cbd4acb perf improvement in mkldnn conv/batch Signed-off-by: ZhenlinLuo <zhenlin.luo@intel.com> 11 April 2018, 01:24:51 UTC
bc780a6 Add Relu Fuse into BatchNorm Signed-off-by: ZhenlinLuo <zhenlin.luo@intel.com> 11 April 2018, 01:21:54 UTC
38286c2 bug fix and perf improvement in mkldnn batch and conv Signed-off-by: ZhenlinLuo <zhenlin.luo@intel.com> 11 April 2018, 01:20:10 UTC
794e65a Add MKLDNN Plus support Signed-off-by: ZhenlinLuo <zhenlin.luo@intel.com> 11 April 2018, 01:14:23 UTC
775ae80 Add Grad update ctrl and inference profiler Signed-off-by: ZhenlinLuo <zhenlin.luo@intel.com> 11 April 2018, 01:11:12 UTC
6e7ff0e Add requirements check for CNTK during Python module import 10 April 2018, 00:16:59 UTC
d3d26d0 Integrate v-rodemo/fix-batchnorm-freedimension into master 09 April 2018, 23:39:20 UTC
82c8c69 Fixing the reader lib bug when number of samples in a chunk is less than the number of workers 08 April 2018, 06:18:46 UTC
d4781c1 updating script to use new learning APIs 06 April 2018, 19:05:45 UTC
0b5b100 Added test for batch norm spatial shape inference 06 April 2018, 16:19:39 UTC
ca0f889 Integrate yuqtang/StageGradOnGatherIndices into master 04 April 2018, 22:30:13 UTC
455aef8 Integrate vadimma/fb into master 04 April 2018, 07:48:31 UTC
0ebaed4 Add an assert for defining serializer 03 April 2018, 20:35:13 UTC
5afcff3 Fix definesMbSize config in bundler 28 March 2018, 22:16:16 UTC
90a41f6 Enabled gather op indcies be computed from parameters. 27 March 2018, 23:48:11 UTC
a7453ed Merge pull request #3065 from Microsoft/szhao/fixsample Update cntk version to 2.5 in the sample 26 March 2018, 22:06:07 UTC
09cdc6c fix the cntk version in the sample 24 March 2018, 10:45:46 UTC
e6cb4c5 Fix spatial batchnorm parameter inference with FreeDimension 23 March 2018, 19:57:45 UTC
5a619b4 Merge branch 'master' into kedeng/mkldnn 23 March 2018, 17:06:47 UTC
64062b1 Rename RandomBuffer to ColMajorBuffer 23 March 2018, 17:05:54 UTC
ed4bd3b Integrate jaliyaek/recur_checkin into master 23 March 2018, 01:33:54 UTC
a5e0912 Fix automated cntk install script for windows 23 March 2018, 00:16:50 UTC
5b04f46 Integrate sptiwari/add_rnn_to_ornn2 into master 22 March 2018, 23:24:40 UTC
8b5dc23 Merge branch 'master' into kedeng/mkldnn 22 March 2018, 22:29:59 UTC
4671bf2 Fixes to MKL-DNN 22 March 2018, 22:29:25 UTC
73c2046 Adding recurrence support to user defined functions. This enables UDF to be called inside recurrent loops. 22 March 2018, 18:44:09 UTC
3f72469 Adding ONNX support for rnnReLu and rnnTanh in OptimizedRNNStack. 22 March 2018, 18:24:32 UTC
65961c9 Integrate liqun/2954 into master 21 March 2018, 22:44:06 UTC
82142f0 Integrate muditj/FixNoRandomizerBug3 into master 21 March 2018, 21:19:55 UTC
aab2567 Integrate liqun/RNNStage into master 21 March 2018, 19:23:28 UTC
c8c02bf Fix no randomizer bug where workers see duplicate data 21 March 2018, 18:57:19 UTC
3e83c56 issue 2954: c# API memory leak 21 March 2018, 18:14:28 UTC
08ba969 support ONNX RNN 21 March 2018, 16:57:54 UTC
3a22962 Integrate sptiwari/mean_var_norm_issue into master 21 March 2018, 05:23:48 UTC
92b7b27 Update .clang-format to remove column limits Otherwise too much code change when formatting with clang-format 21 March 2018, 01:57:47 UTC
7068329 Fixes to mean_variance_normalization and ONNX LSTM to run Office models. 20 March 2018, 17:53:06 UTC
0254c82 Fix GPU BrainScript 20 March 2018, 01:15:33 UTC
db0d8ab Merge branch 'master' into kedeng/mkldnn 19 March 2018, 22:20:14 UTC
c361798 Fix GPU build 19 March 2018, 22:19:34 UTC
10c327e No aggregation or NCCL initialization if only one node used undef _SCL_SECURE_NO_WARNINGS for training 19 March 2018, 16:54:41 UTC
542406d No need to aggregate with one worker in brainscript Remove msg printint GPU UUID 19 March 2018, 16:50:09 UTC
back to top