https://github.com/Microsoft/CNTK

sort by:
Revision Author Date Message Commit Date
dcdad16 Introduce latticeConfigPath to SE node 14 February 2018, 17:57:42 UTC
9d2cda7 Some refactoring 14 February 2018, 06:11:47 UTC
d465167 Introduce lattice reader verbosity 14 February 2018, 06:11:47 UTC
8ace52b Merge branch 'master' of https://github.com/Microsoft/CNTK 14 February 2018, 06:11:16 UTC
aa7447e First round of changes for ONNX FreeDimension support. 13 February 2018, 19:32:21 UTC
0981d92 Merge branch 'master' of https://github.com/Microsoft/CNTK 13 February 2018, 18:36:19 UTC
19719a6 Fast tensor ops using MKL - Accelerates some common tensor ops in Intel CPU inference for float32, especially for fully connected networks - Can be turned on/off by cntk.cntk_py.enable_cpueval_optimization()/cntk.cntk_py.disable_cpueval_optimization() 09 February 2018, 06:22:27 UTC
08bcbdc Merge pull request #2690 from MingyuJoo/CNTK-2-PretrainedModels Adding ResNet101_ImageNet_CNTK/ResNet152_ImageNet_CNTK in download_model.py. 09 February 2018, 00:42:08 UTC
bd121f6 Merge pull request #2843 from kaiidams/kaiidams/typotut206 Fixed typos in tutorials. 09 February 2018, 00:33:47 UTC
8b71ff0 Merge pull request #2881 from zingdle/patch-2 Remove redundant code in CNTK_204_Sequence_To_Sequence.ipynb 09 February 2018, 00:30:07 UTC
2b94e7d Merge pull request #2880 from rnrneverdies/master Fix TensorBoard on_write_test_summary method 09 February 2018, 00:27:18 UTC
06ee7b8 Merge pull request #2940 from kazuyaujihara/patch/csexampletypo Fix typo in C# examples 09 February 2018, 00:25:21 UTC
062ce4d Merge pull request #2822 from tsbertalan/master Grammar fix. 09 February 2018, 00:24:11 UTC
08ba214 Update README.md 09 February 2018, 00:23:39 UTC
1142563 Integrate sptiwari/onnx_scalar_constant into master 08 February 2018, 05:15:07 UTC
aa1aaf8 Bug fix in ONNX broadcasting for scalar. 08 February 2018, 01:37:34 UTC
c5abe7f Add KERA_BACKGROUND=cntk into CNTK Dockerfiles 07 February 2018, 18:34:51 UTC
56e1f3b Fix typo 07 February 2018, 14:34:48 UTC
83b1c9b Merge branch 'vadimma/LatticeParallel' of https://github.com/Microsoft/CNTK 06 February 2018, 22:33:21 UTC
4961226 Make the lattice deserialization parallel 06 February 2018, 22:21:32 UTC
ceaa223 some chnage 06 February 2018, 21:29:50 UTC
425baf4 Merge branch 'master' of https://github.com/Microsoft/CNTK 06 February 2018, 21:08:18 UTC
d615602 Integrate sptiwari/onnx_leakyrelu_fix into master 06 February 2018, 08:05:23 UTC
6556e1f Passin device ID in DistComm for NCCL initilization 05 February 2018, 21:50:39 UTC
977e402 Backward compat bug fix in LeakyReLu reverting alph to type double. 05 February 2018, 19:08:40 UTC
5fd5bf8 Make the lattice deserialization parallel 02 February 2018, 06:34:51 UTC
1c578d3 Merge branch 'master' of https://github.com/Microsoft/CNTK 02 February 2018, 05:55:25 UTC
3660b7a Node timing and profile details format in chrome://tracing. Working example in ./Examples/Image/Classification/MLP/Python/SimpleMNIST.py Note that node timing would be added to profiler details when profiler is enabled, i.e. import cntk as C C.debugging.debug.set_node_timing(True) C.debugging.start_profiler() C.debugging.enable_profiler() trainer|evaluator|function executions trainer|evaluator|function.print_node_timing() C.debugging.stop_profiler() 02 February 2018, 05:53:46 UTC
c3f7738 change isnan issue 02 February 2018, 05:17:33 UTC
3a19d2f Adding find_by_uid to logging API. 01 February 2018, 22:58:37 UTC
fbbbaa9 Merge release/2.4 Squashed commit of the following: commit 82e025bbe319008c535741a9fc78367032cc3be0 Merge: 0702c09c3 ed5451c83 Author: Manik Jindal <manikj@microsoft.com> Date: Wed Jan 31 21:23:23 2018 -0800 Merge branch 'release/2.4' into manikj/merge-release-2.4 commit 0702c09c3c3cc1ae1e4cba28df3ce64f091a8c3e Merge: 32f6844ca 1afe40461 Author: Manik Jindal <manikj@microsoft.com> Date: Wed Jan 31 21:23:11 2018 -0800 Merge branch 'release/2.4' (early part) into manikj/merge-release-2.4 commit 32f6844ca84797db2dac8959eeb72576692986d0 Merge: 42a47a326 1f15d4649 Author: Manik Jindal <manikj@microsoft.com> Date: Wed Jan 31 21:21:55 2018 -0800 Merge branch 'release/2.4' (early part) into manikj/merge-release-2.4 commit ed5451c83416c5084dedf1a684f03653fe5863ee Author: Manik Jindal <manikj@microsoft.com> Date: Wed Jan 31 19:32:04 2018 -0800 MakeBinaryDrop: Fix lib names commit c528a56d4a1fa12f967e4bbab022a0d06b5704ea Author: Manik Jindal <manikj@microsoft.com> Date: Wed Jan 31 12:35:40 2018 -0800 Update README.md commit 490a58e4a9cf01d9114373b9444f0fbaf055581b Author: Manik Jindal <manikj@microsoft.com> Date: Wed Jan 31 12:08:57 2018 -0800 Add Halide to Release notes commit 2196299a71325b0959dc626ce548035db0634e58 Author: Manik Jindal <manikj@microsoft.com> Date: Wed Jan 31 11:48:45 2018 -0800 MakeBinaryDrop: Change Cuda libs to 9.0 commit 3df48ddd3ac38c24c5aa25bdb9b77d98116a3d8d Author: Manik Jindal <manikj@microsoft.com> Date: Wed Jan 31 06:06:29 2018 -0800 Add release notes commit 1afe404611f1f9a7cde0f43923382101cf15b823 Author: Manik Jindal <manikj@microsoft.com> Date: Wed Jan 31 06:05:05 2018 -0800 Replace 2.4+ with 2.4 commit 1f15d464960956a8192f137ebedf2753a36d83cc Author: Manik Jindal <manikj@microsoft.com> Date: Wed Jan 31 06:04:24 2018 -0800 Bump version to 2.4 01 February 2018, 05:38:52 UTC
42a47a3 Integrate yacheo/h-softmax into master 01 February 2018, 05:07:31 UTC
e129bcf MakeBinaryDrop: Change Cuda libs to 9.0 01 February 2018, 02:48:12 UTC
7877e3f add hierarchical softmax 01 February 2018, 02:22:27 UTC
abfc3a8 Update current_iteration.md 31 January 2018, 19:44:29 UTC
688b96e Update current_iteration.md 31 January 2018, 19:26:52 UTC
b09ff2b Merge branch 'master' into liqun/onnx17Stage2 31 January 2018, 13:50:43 UTC
4bb3031 Making axis attribute optional in Concat. 31 January 2018, 08:52:52 UTC
22cea7c broadcast of scalar 31 January 2018, 04:48:15 UTC
e91df39 Pooling missing code fix 30 January 2018, 22:14:36 UTC
f18bcda Updating release notes with ONNX and other feature work (e.g. group convolution). 30 January 2018, 18:05:47 UTC
2a764a7 add ImageScaler, fix ConvTranspose 30 January 2018, 16:46:06 UTC
00158ff Merge pull request #2887 from Microsoft/marhamil-fix-jar Update post-build.cmd 30 January 2018, 03:13:01 UTC
6ea2c67 Break Nuget packages Separate out dependencies from Nuget packages to reduce size of each individual package because Nuget service do not allow to upload a package of size more than 333MB. Current list of packages: CNTK.Deps.Cuda - Cuda dlls for GPU package CNTK.Deps.MKL - MKL dlls CNTK.Deps.OpenCV.Zip - OpenCV and Zip dlls CNTK.CPUOnly - CNTK CPU only Nuget package. Dependencies: MKL, OpenCV.Zip CNTK.GPU - CNTK GPU Nuget package. Dependencies: Cuda, MKL, OpenCV.Zip CNTK.UWP.CPUOnly - CNTK CPU Nuget package for UWP. Microsoft.Research.CNTK.CpuEval-mkl - Legacy CNTK eval Nuget package. Dependencies: MKL 30 January 2018, 02:35:28 UTC
6e846aa fix makefile 29 January 2018, 20:44:29 UTC
8569868 fix variable 29 January 2018, 18:53:32 UTC
166c8bc add debug dll support 29 January 2018, 16:12:37 UTC
6db8b6e Add image writer SOs and zlib dlls 29 January 2018, 15:58:34 UTC
08cc45c Edit makefile 29 January 2018, 15:56:46 UTC
be931c8 Update post-build.cmd 29 January 2018, 15:56:45 UTC
ba9c2e7 Fix build if optional MKLDNN is not present 28 January 2018, 19:27:32 UTC
950ac47 Integrate jaliyaek/config into master 28 January 2018, 12:25:33 UTC
dd47e21 Making Halide default in Jenkin builds 27 January 2018, 01:17:58 UTC
343f383 Remove Python 3.4 support 26 January 2018, 22:55:22 UTC
5edcb78 Add invalid lattice handling 26 January 2018, 22:47:30 UTC
ce276be debug2 26 January 2018, 17:22:27 UTC
6d9fcf5 debug code 26 January 2018, 05:59:17 UTC
2c90f08 Test setting gradint to zero for problematic inputs 26 January 2018, 04:29:33 UTC
ffc7507 Integrate jaliyaek/halide_squash2 into master 25 January 2018, 07:14:14 UTC
32c4e04 Update CNTK_204_Sequence_To_Sequence.ipynb remove redundant code. create_reader() is correctly defined in block 10. 25 January 2018, 04:10:16 UTC
af26e4f Fix TensorBoard on_write_test_summary method (#2878) 25 January 2018, 03:42:58 UTC
a7a52d7 Adding halide based binary convolution operators and its dependancies 25 January 2018, 00:41:17 UTC
851ea5d Fix linux binary drop script for nccl 24 January 2018, 06:18:32 UTC
387725d Merge branch 'kedeng/fixDocker2' 24 January 2018, 03:28:47 UTC
9f6cced Fix Dockerfiles under Tools/docker 23 January 2018, 21:44:01 UTC
a087490 Integrate sptiwari/remove_find_by_uid into master 23 January 2018, 20:16:49 UTC
a40285c Rename default Anaconda python environment Depending on the CNTK setup steps followed, either cntk-py<version> or <cntkdev-py<version> is installed. cntk-py<version> will be adopted as the default from now on. 23 January 2018, 18:43:51 UTC
e883520 Remove redefinition of CNTK_COMPONENT_VERSION Math.vcsproj inherits this macro from CNTK.Common.props already 23 January 2018, 18:28:53 UTC
3cf3af5 CNTK support for CUDA 9 CNTK now supports CUDA 9/cuDNN 7. This requires an update to build environment to Ubuntu 16/GCC 5 for Linux, and Visual Studio 2017/VCTools 14.11 for Windows. With CUDA 9, CNTK also added a preview for 16-bit floating point (a.k.a FP16) computation. Please check out the example of FP16 in ResNet50 at /Examples/Image/Classification/ResNet/Python/TrainResNet_ImageNet_Distributed.py Notes on FP16 preview: * FP16 implementation on CPU is not optimized, and it's not supposed to be used in CPU inference directly. User needs to convert the model to 32-bit floating point before running on CPU. * Loss/Criterion for FP16 training needs to be 32bit for accumulation without overflow, using cast function. Please check the example above. * Readers do not have FP16 output unless using numpy to feed data, cast from FP32 to FP16 is needed. Please check the example above. * FP16 gradient aggregation is currently only implemented on GPU using NCCL2. Distributed training with FP16 with MPI is not supported. * FP16 math is a subset of current FP32 implementation. Some model may get Feature Not Implemented exception using FP16. * FP16 is currently not supported in BrainScript. Please use Python for FP16. To setup build and runtime environment on Windows: * Install [Visual Studio 2017](https://www.visualstudio.com/downloads/) with following workloads and components. From command line (use Community version installer as example): vs_community.exe --add Microsoft.VisualStudio.Workload.NativeDesktop --add Microsoft.VisualStudio.Workload.ManagedDesktop --add Microsoft.VisualStudio.Workload.Universal --add Microsoft.Component.PythonTools --add Microsoft.VisualStudio.Component.VC.Tools.14.11 * Install [NVidia CUDA 9](https://developer.nvidia.com/cuda-90-download-archive?target_os=Windows&target_arch=x86_64) * From PowerShell, run: /Tools/devInstall/Windows/DevInstall.ps1 * Start VCTools 14.11 command line, run: cmd /k "%VS2017INSTALLDIR%\VC\Auxiliary\Build\vcvarsall.bat" x64 --vcvars_ver=14.11 * Open /CNTK.sln from the VCTools 14.11 command line. Note that starting CNTK.sln other than VCTools 14.11 command line, would causes CUDA 9 [build error](https://developercommunity.visualstudio.com/content/problem/163758/vs-2017-155-doesnt-support-cuda-9.html). To setup build and runtime environment on Linux using docker, please build Unbuntu 16.04 docker image using Dockerfiles /Tools/docker. For other Linux systems, please refer to the Dockerfiles to setup dependent libraries for CNTK. 23 January 2018, 00:58:56 UTC
8b6126c Refactor hmm init 22 January 2018, 21:35:09 UTC
4841c79 Introduce exception handling for numberic instability of SE training 22 January 2018, 20:28:04 UTC
eeb0645 Removing find_by_uid API. 22 January 2018, 18:33:41 UTC
3765da9 Support CNTK installation in VirtualEnv CNTK installation using pip place dlls in the root folder of the python environment. Unlike Conda, VirtualEnv doesn't add this folder in library search path and hence 'import CNTK' is broken in VirtualEnv. Fix: Place CNTK libraries in site-packages/CNTK/lib and add the same to library search path. 20 January 2018, 01:59:03 UTC
997d3a1 Replace CntkCsAssemblyVersion by CntkComponentVersion on VS projects 19 January 2018, 22:36:34 UTC
f36c73c Integrate sptiwari/find_by_uid4 into master 19 January 2018, 19:41:04 UTC
1c86a62 Set default verbosity to 0 19 January 2018, 02:20:11 UTC
4836687 Fixing bug in ONNX pooling node serialization. 17 January 2018, 23:10:23 UTC
f07a98d Adding find_by_uid to logging API. 17 January 2018, 22:35:23 UTC
8066621 Fixing BPTT for case when the minibatch size changes 17 January 2018, 01:20:23 UTC
c3d01e4 Enabling evaluation on GPU 17 January 2018, 01:20:23 UTC
5bdaed7 Integrate yuqtang/TimesOnFreeAxes into master 16 January 2018, 23:27:02 UTC
10d7130 Integrate vadimma/LatticeSerializer into master 16 January 2018, 22:39:23 UTC
8b6af80 Update the max_sequence_length 16 January 2018, 03:41:23 UTC
05cbd0a Remove too much logging 16 January 2018, 00:40:06 UTC
e6dc1a4 Refactor lattice sequence initialization' 15 January 2018, 23:22:11 UTC
01c1509 Fixed typos in tutorials. CNTK_206A_Basic_GAN.ipynb CNTK_206B_DCGAN.ipynb CNTK_206C_WGAN_LSGAN.ipynb 15 January 2018, 12:39:06 UTC
77c95a4 Trim lattice TOC line entries 15 January 2018, 08:08:02 UTC
d929ecd Add logs for lattice index builder 15 January 2018, 07:45:18 UTC
1a81d41 Fix batch matmul test failures 15 January 2018, 07:32:12 UTC
d26f547 debug large lattice entries 15 January 2018, 04:57:11 UTC
1fe17f8 Add some debugging statements 15 January 2018, 01:13:02 UTC
f79324c Ensure the lattice serializer handles several chunks in a file 15 January 2018, 00:17:15 UTC
ea6834b Restrict the size of lattice 14 January 2018, 17:57:29 UTC
5962f71 Introduce the maxSequenceSize config to the HTK feature reader 14 January 2018, 06:54:37 UTC
7d67daf Enable max seq size config in HTK reader 14 January 2018, 00:16:40 UTC
15e705d add batch matmul 13 January 2018, 02:41:34 UTC
01c22ad Change the docstring 13 January 2018, 01:38:02 UTC
bcbf95b Some fixes after rebase 12 January 2018, 20:07:47 UTC
f3e9993 Remove main from the lattice test 12 January 2018, 19:06:41 UTC
2cac661 Update the python test 12 January 2018, 19:06:41 UTC
back to top