https://github.com/Microsoft/CNTK

sort by:
Revision Author Date Message Commit Date
5dffd71 Added DSSM tutorial 16 October 2017, 20:25:19 UTC
8e36d30 rename tmpdir 13 October 2017, 14:55:08 UTC
00ced21 use CNTKNativeUtils load in java 13 October 2017, 14:55:08 UTC
1280bde make java so variable names in makefile less confusing 13 October 2017, 14:55:08 UTC
4c56af0 cudnn64_5 -> cudnn64_6 13 October 2017, 14:55:08 UTC
1a8a839 add cuda deps 13 October 2017, 14:55:08 UTC
e0c0ec7 fix sln, and make post-build give better message 13 October 2017, 14:55:07 UTC
78ee54d use correct mkl lib path for jar 13 October 2017, 14:55:07 UTC
0d1834f add DLL's and so's to jar (squashed) minor fixes fixing Native utils path, loader, and the Native Load Utils remove path and classpath setting in tests cleaning up native utils Unified CNTKNativeUtils class. Changed the code generation to use CNTKNativeUtils directly instead of the intermediary CNTK.init(). adding fixes to post bild Added NATIVE_LOAD_MANIFEST for the cases when only specific high-level libraries should be loaded linux side Add gpu support linux gpu .so copying 13 October 2017, 14:55:07 UTC
a631e37 Integrate lakshayg-atanh into master 13 October 2017, 05:52:19 UTC
bbe8605 Fix typos (#2470) 13 October 2017, 05:00:40 UTC
02b3640 Merge branch 'atanh' of https://github.com/lakshayg/CNTK into lakshayg-atanh Atanh PR... 12 October 2017, 23:40:18 UTC
d2fcb7d change return RuntimeError to raise RuntimeError 12 October 2017, 14:19:00 UTC
75a7cbe SimpleDistAggregation: Donot copy data from GPU to CPU when NCCL is used 11 October 2017, 23:53:17 UTC
8a61d67 fix combine java impl The line inside the loop in the java idiomatic wrapper refers to the wrong variable. 11 October 2017, 17:53:36 UTC
84030c5 Integrate ebarsoum/lrn_doc into master 10 October 2017, 23:34:09 UTC
8aa7466 Update README.md 10 October 2017, 20:53:40 UTC
d011e14 Update README.md 10 October 2017, 20:53:01 UTC
302b9db Put back LRN documentation. 10 October 2017, 17:37:09 UTC
d57ad1d Adding ONNX format support to CNTK. 10 October 2017, 08:01:43 UTC
0b384cb Crosstalk Caffe Bug Fixes Resolved bugs due to: 1. Scaling in batch normalization was done twice 2. Splicing did not take into account a default axis 3. Did not support newer protobuf versions. 4. Stdout was not being ignored during validation 09 October 2017, 17:10:15 UTC
08e367e Expose distributed evaluation in Python 05 October 2017, 23:33:39 UTC
2f9e3e8 Update baseline after disabling Index_64MB_with_caching_check_perf test 05 October 2017, 21:25:59 UTC
c8e55c0 Disable Index_64MB_with_caching_check_perf test because it's flaky on Jenkins. Will need further investigation later. 05 October 2017, 15:07:26 UTC
b184e73 Bug fix. 04 October 2017, 22:52:09 UTC
a1f5d75 Revise test to disable adam, fsadagrad and rmsprop when running on cpu. 04 October 2017, 17:03:47 UTC
cdd62cf Faster adadelta for sparse gradients added unit test and improved parallelism cleanup bug fixes and CR feedback 04 October 2017, 17:03:40 UTC
0d13098 Fix typos (#2434) 04 October 2017, 05:00:27 UTC
b224a7d fix rest issues find with code scan 29 September 2017, 02:01:29 UTC
af149ee Refactoring of general graph algorithms from ComputationNetwork 28 September 2017, 10:32:57 UTC
7243389 Merge branch 'master' into liqun/ArgFix 28 September 2017, 05:16:26 UTC
3da3f5a extra va_end_start pair 27 September 2017, 21:25:49 UTC
bded222 Corrected docstring of Adam optimizer. 27 September 2017, 16:59:17 UTC
fceab10 Integrate juncgu/optimizingAgg into master 27 September 2017, 09:12:56 UTC
f43c6bf Integrate kedeng/issue2388 into master 27 September 2017, 08:13:19 UTC
79e0039 Fix to build on CPU fix build on Windows 27 September 2017, 02:46:18 UTC
afc8deb optimize AggregateGradientsImpl (async pipeline) for non-GDR & GPU scenario rename "j" to currentGradientIndex in line321 update description in line309; rename j to currentGradientIndex fix bug (from Junjie) in new gradient aggregator pipeline (== -> !=) move first D-to-H copy out of the loop define "size_t i" in simple gradient aggregator 27 September 2017, 02:46:16 UTC
62a80d2 Fix Issue #2388 27 September 2017, 00:06:33 UTC
40162ba faster way to get the logits in the nce docstring 26 September 2017, 18:46:50 UTC
6120812 Update README.md 25 September 2017, 22:55:13 UTC
a6302a9 push 25 September 2017, 22:01:57 UTC
dfef3d6 Avoiding segmentation fault while introducing udf in recurrent loop 25 September 2017, 11:59:11 UTC
474d468 minor bug fix (#2402) 24 September 2017, 12:36:25 UTC
1cbccd0 Merge branch 'master' into ebarsoum/dilation_with_padding Merge with master... 21 September 2017, 20:03:28 UTC
99d2c82 Pick up submodule fixes for block momentum in python and add test Remove some dead code 21 September 2017, 06:14:16 UTC
3bf78e8 Fix dilation convolution bug with auto padding 21 September 2017, 00:12:38 UTC
ec8d6d6 Update the CTC tutorial further 19 September 2017, 19:15:40 UTC
7fb2acf update CTC tutorial 19 September 2017, 19:15:40 UTC
0c6ebba Adding group convolution. 19 September 2017, 00:14:47 UTC
c3e4f66 Tests/Install/linux/prep-run-test.sh: do not run CNTK_104_Finance_Timeseries_Basic_with_Pandas_Numpy.ipynb 18 September 2017, 17:59:05 UTC
96cdd12 Bump version numbers 18 September 2017, 17:58:47 UTC
bced9cf remove inclusion of dependency dlls, as it has been referenced in UWP nuget package. 18 September 2017, 17:58:31 UTC
e4cc788 Tests/EndToEndTests: Linux: test green for Tesla M60 18 September 2017, 09:04:32 UTC
83dab7c Fix typos (#2370) This PR fixes some typos: CNKT 18 September 2017, 05:11:59 UTC
75fa318 Fix typos (#2212) 17 September 2017, 19:38:46 UTC
cd6748f Handfew typos fixed (#2366) 17 September 2017, 19:37:41 UTC
4a5f64c Fix typos (#2338) This PR fixes some typos: CTNK 17 September 2017, 01:08:08 UTC
159a611 Minor fixes (proposal) (#2357) plus typo? 17 September 2017, 01:07:39 UTC
6cc4451 Fixes "'exp' was not declared" (#2364) This fixes the error: "error: ‘exp’ was not declared in this scope return momTC == 0.0 ? 0 : exp(-1.0 / momTC);",when compiling CNTK with GCC. Reference: https://github.com/Microsoft/CNTK/issues/2341 17 September 2017, 01:06:26 UTC
f275a54 Update README.md 16 September 2017, 12:11:58 UTC
2f9e469 Added sequence classification to current language understanding tutorial 15 September 2017, 22:50:17 UTC
bfe1d3c Tests/EndToEndTests: add some CC5 baselines for Linux 15 September 2017, 10:54:23 UTC
98a6b42 implement atanh operation 15 September 2017, 10:47:44 UTC
741d01e language edits 15 September 2017, 06:14:51 UTC
ef49c98 Integrate yuqtang/StageManualUseLearners2 into master 15 September 2017, 06:08:38 UTC
c525d71 Integrate yuqtang/StageCleanStartupPrints2 into master 15 September 2017, 00:58:42 UTC
26fa020 Squashed commit of the following: Updated the manual on how to use learners to explain the new Learner APIs for release 2.2. 14 September 2017, 23:23:44 UTC
0f3033d Integrate yuqtang/StagingFixNewLearner2 into master 14 September 2017, 22:55:00 UTC
d49774d Corrected the case of the word "version". 14 September 2017, 22:19:29 UTC
8c87d1a Update release notes. 14 September 2017, 22:19:14 UTC
1e32914 Squashed commit of the following: Removed sensitive information in CNTK start-up message. 14 September 2017, 22:12:17 UTC
a01f8cd Squashed commit of the following: Add python side work-around to allow the combination of mean gradient and specifying number samples with the right implementation; in Python, achieve the legacy logic of having per sample learning rate with mean gradient by overriding the per sample learning rate with ignored_minibatch_size. Also clarification in docstring. 14 September 2017, 20:01:55 UTC
e48892c Merge branch 'master' of https://github.com/Microsoft/cntk 14 September 2017, 18:11:34 UTC
0eb7a1f update keras and tensorboard text 14 September 2017, 18:10:55 UTC
d9b0da2 Integrate liqun/iterationnote2 into master 14 September 2017, 06:08:59 UTC
484a4f7 style fix 14 September 2017, 05:58:15 UTC
10fb523 CSharp iteration note 14 September 2017, 05:42:12 UTC
f27ed0a Merge branch 'master' of https://github.com/Microsoft/cntk 14 September 2017, 05:10:27 UTC
b4a1704 update contributor list. 14 September 2017, 05:08:47 UTC
035d79f Integrate ebarsoum/iteration into master 14 September 2017, 02:03:38 UTC
9ccbc67 Merge branch 'master' of https://github.com/Microsoft/cntk 14 September 2017, 01:34:28 UTC
66b996f update with keras and opencv info 14 September 2017, 01:34:20 UTC
d007293 Merge with master. 14 September 2017, 01:29:10 UTC
5d9369e format change 14 September 2017, 01:21:00 UTC
e31a11a Add acknowledgments 14 September 2017, 01:13:07 UTC
e16816e More style update 14 September 2017, 01:07:38 UTC
99adb85 Fix style in current_iteration.md 14 September 2017, 01:01:06 UTC
d5cf5e1 current_iteration.md update on NCCL2, NVidia FP16 and Intel MKL works 14 September 2017, 00:57:02 UTC
7dd2ce1 Update iteration planning. 14 September 2017, 00:34:15 UTC
4dfd1e2 fix version 13 September 2017, 22:28:50 UTC
78f122c add more text 13 September 2017, 22:17:50 UTC
adccabc Update LICENSE.md changed license verbiage for Intel MKL to allow us to move up to MKL 2017 13 September 2017, 21:58:56 UTC
6c2f2a1 Integrate zhouwang/releasenote-update into master 13 September 2017, 21:34:33 UTC
6f5f285 Adding links to object detection tutorial/demos in 2.2 release notes. 13 September 2017, 21:23:36 UTC
7ca42da add reference to doc page 13 September 2017, 21:19:13 UTC
162c7c5 Removing semantic segmentation from 2.2 release note. 13 September 2017, 21:10:33 UTC
4a875a1 Updated FasterRCNN and convolution FreeDimension bullets in 2.2 release note. 13 September 2017, 21:07:25 UTC
1797ff3 Integrate sptiwari/fastrcnn_readme into master 13 September 2017, 20:53:37 UTC
bcfbfbc Added links to the manuals as well. 13 September 2017, 20:46:17 UTC
0e24a13 Added link to python doc site. 13 September 2017, 20:40:51 UTC
back to top