5fc6c2a | Peyman Manikashani | 30 November 2018, 19:22:25 UTC | sequence_isfirst_islast test fix for python 2.7 | 30 November 2018, 19:22:25 UTC |
75b0141 | Peyman Manikashani | 28 November 2018, 00:47:28 UTC | Adding support for exporting CNTK's Sequence::IsFirst and Sequence::IsLast nodes | 29 November 2018, 23:57:43 UTC |
d1d1133 | Bowen Bao | 13 November 2018, 19:46:59 UTC | ConvTranspose asymmetric padding * Temporarily reverse the extra padding location in case of SAME_UPPER vs SAME_LOWER for convTranspose to match with onnxruntime. * In case of importing asymmetric padding convTranspose, use symmetric pads by alter the output_shape and pads, and attach a slice node afterwards to enable cudnn. * Fix a bug in slice/squeeze attribute axes export. | 27 November 2018, 17:33:15 UTC |
196df1a | Emad Barsoum | 26 November 2018, 20:49:56 UTC | Update image layout in readme. | 26 November 2018, 20:49:56 UTC |
eeae2bb | Emad Barsoum | 26 November 2018, 19:57:39 UTC | Add netron news and update iteration. | 26 November 2018, 19:57:39 UTC |
8c15683 | Vadim Mazalov | 24 November 2018, 09:12:41 UTC | Merge branch 'vadimma/binmlf' | 24 November 2018, 09:12:41 UTC |
0f1b4e0 | Sergii Dymchenko | 24 November 2018, 03:44:06 UTC | Merge pull request #3512 from changeworld/patch-1 Fix typo | 24 November 2018, 03:44:06 UTC |
b566e23 | Takashi Takebayashi | 23 November 2018, 15:25:07 UTC | Fix typo Ture -> True | 23 November 2018, 15:25:07 UTC |
0c1d283 | liqfu | 22 November 2018, 16:56:07 UTC | update latest onnx and onnxruntime, fix shape inference | 22 November 2018, 16:56:07 UTC |
631392e | Vadim Mazalov | 15 November 2018, 03:21:11 UTC | Update bin mlf reader | 15 November 2018, 15:24:40 UTC |
d7101a2 | Vadim Mazalov | 12 November 2018, 20:34:56 UTC | Clean up bin mlf test | 12 November 2018, 20:34:56 UTC |
8d87948 | Vadim Mazalov | 12 November 2018, 20:19:37 UTC | Ensure mlf reader is included in the test | 12 November 2018, 20:19:37 UTC |
72200ca | Vadim Mazalov | 11 November 2018, 22:29:27 UTC | Add bin mlf reader tests | 11 November 2018, 22:29:27 UTC |
7ef57de | Vadim Mazalov | 11 November 2018, 05:32:28 UTC | Add dim label to bin mlf reader | 11 November 2018, 05:32:28 UTC |
22c7c3c | Aghagolzadeh | 08 November 2018, 22:32:13 UTC | Update BlockMomentumDistributedLearner.h | 08 November 2018, 22:32:13 UTC |
69df29e | Liqun Fu | 08 November 2018, 22:14:44 UTC | Merge branch 'liqun/seqopsStage' | 08 November 2018, 22:14:44 UTC |
2405457 | Aghagolzadeh | 08 November 2018, 20:27:55 UTC | Update BlockMomentumDistributedLearner.h | 08 November 2018, 20:27:55 UTC |
ab4bee2 | liqfu | 08 November 2018, 02:36:20 UTC | Support RNN ops in a Scan loop Update with latest ONNX Update with latest ONNX graph IR Support sequence ops - Sequence::Gather, Sequence::PastValue, Sequence::FutureValue, etc. | 08 November 2018, 02:36:20 UTC |
b51e8c2 | Vadim Mazalov | 06 November 2018, 00:32:59 UTC | Expose binmlf in python | 06 November 2018, 00:32:59 UTC |
3f46cf0 | Bowen Bao | 26 October 2018, 00:50:19 UTC | Updates on several ONNX exports. * ConvTranspose outputShape: now pads values are always exported even when outputShape is given. The reason is that CNTK and ONNX have different padding specs. * Flatten: in CNTK flatten does not affect batch axis, this should be preserved in ONNX. | 03 November 2018, 00:18:07 UTC |
a19ce9e | Bowen Bao | 02 November 2018, 17:43:24 UTC | Resolve forward_backward docstring Sphinx error. | 02 November 2018, 17:43:24 UTC |
29818ff | Bowen Bao | 01 November 2018, 06:58:24 UTC | Merge branch 'bowbao/onnxruntime_ci_stage2' | 01 November 2018, 06:58:24 UTC |
0cd2fae | Bowen Bao | 31 October 2018, 23:43:42 UTC | Merge branch 'pull/3476' | 31 October 2018, 23:43:42 UTC |
fca1396 | Bowen Bao | 18 October 2018, 23:42:36 UTC | Add onnx_test_runner verification in CI. * onnx_test_runner.exe will be called on win64 GPU tests to verify if the output data produced by CNTK in onnx_op_test and onnx_model_test(cntk_model_test) matches in onnxruntime. | 31 October 2018, 22:46:56 UTC |
a0d6f14 | delzac | 31 October 2018, 13:05:27 UTC | Fixed type and error in identifying the deserialiser | 31 October 2018, 13:05:27 UTC |
cdd18d2 | Bowen Bao | 29 October 2018, 18:16:42 UTC | Update LSTM/Truncated-Kaldi CPU baseline | 29 October 2018, 18:16:42 UTC |
3cef1fd | liqfu | 28 October 2018, 17:22:07 UTC | fix python 27 onnx test | 28 October 2018, 17:22:07 UTC |
1ff5548 | Bowen Bao | 26 October 2018, 17:40:01 UTC | Set linux pip install future version to 0.16.0 * This was not set explicitly. Lately the version was updated to 0.17.0 and is causing failure in CI. | 26 October 2018, 17:40:01 UTC |
0b41602 | Bowen Bao | 23 October 2018, 17:50:56 UTC | Update more baseline affected by fixing InvStdDev. | 26 October 2018, 00:43:52 UTC |
77dad35 | delzac | 25 October 2018, 11:17:25 UTC | Added better documentation to forward_backward | 25 October 2018, 11:17:25 UTC |
e940605 | liqfu | 20 October 2018, 04:36:21 UTC | Support ONNX Scan op | 20 October 2018, 04:36:21 UTC |
a55e871 | Bowen Bao | 18 October 2018, 22:23:03 UTC | Fix InvStdDev. * Issue was that AssignSqrOfDifferenceOf(beta, input, mean, alpha) assigns mean value to the gaps in input. These values are then reduced within this function, leading to incorrect results. The fix is to execute assign and reduce separately, and mask gaps to zero again before reducing. * Update test baseline affected by this change (err is lowered by <1%). | 19 October 2018, 17:24:11 UTC |
0ffdcf7 | Bowen Bao | 04 October 2018, 01:33:40 UTC | Overhaul node name export & other fixes * Overhaul node name export. Create static class UniqueNodeNameStorage to manage ONNX node name generation with maintained one-to-one mapping between CNTK Uid, while preserving the original CNTK node name by best efforts (#3358). * Update onnx_op_test to test the preservation of original CNTK node names in exported/imported models. * Update onnx_test_helper to support proper linking of test data and onnx model input/output with unique names. * Update onnx_test_helper to generate .bat file to run exported models in further onnxruntime verification. * Fix Sum import to support arbitrary number of inputs. Sum implementation in CNTK backend is loop of Plus, which takes care of potential broadcast issues. | 18 October 2018, 01:36:48 UTC |
da2e610 | Yang Chen | 12 October 2018, 21:12:23 UTC | Replaced wchar/wstring with char/string in C interface | 12 October 2018, 21:12:23 UTC |
149d87b | Spandan Tiwari | 05 October 2018, 21:06:56 UTC | Adding ONNX export support for OneHotOp. | 05 October 2018, 21:06:56 UTC |
6063866 | Bowen Bao | 02 October 2018, 03:41:06 UTC | Merge branch 'bowbao/pooling_pad_offset' | 02 October 2018, 03:41:06 UTC |
bf37aad | Bowen Bao | 01 October 2018, 20:50:24 UTC | Fix pad offset computation for pooling * Compute keyInterior according to the updated algorithm for computing cell offset key. * Update unittest of avg_pooling/max_pooling for cases that requires auto_padding = True. Previous test cases cover only those that do not need padding. | 02 October 2018, 00:21:30 UTC |
7c838d3 | TJ | 01 October 2018, 22:44:06 UTC | fix csharp examples and references | 01 October 2018, 22:45:13 UTC |
fcdeef6 | Bowen Bao | 26 September 2018, 22:53:36 UTC | Support crop_manual export & import. | 29 September 2018, 20:46:17 UTC |
a36fae8 | Bowen Bao | 21 September 2018, 23:04:14 UTC | Support logPlus(log_add_exp) export to ONNX * ONNX supports similar op ReduceLogSumExp. Conversions are added when exporting. * Refactored CNTKToONNXHelper::BroadcastInputsIfNeeded to support more generalized cases. | 28 September 2018, 22:59:55 UTC |
c2072cc | Spandan Tiwari | 27 September 2018, 17:25:11 UTC | Add support for ONNX export of StraightThrough op. | 27 September 2018, 17:25:11 UTC |
1aab76a | Spandan Tiwari | 27 September 2018, 01:01:34 UTC | Updating ONNX submodule hash to include defs for ConstantLike and EyeLike ops. | 27 September 2018, 01:01:34 UTC |
ce503f8 | Peyman Manikashani | 26 September 2018, 00:09:23 UTC | pooling export fix for backward compatibility | 26 September 2018, 00:09:39 UTC |
9165fd0 | Ke Deng | 25 September 2018, 00:26:34 UTC | Merge branch 'kedeng/fixCrash' | 25 September 2018, 00:26:34 UTC |
58f810f | liqfu | 22 September 2018, 16:53:27 UTC | update with ONNX1.3 and latest onnxruntime | 22 September 2018, 16:53:27 UTC |
1489de8 | KeDengMS | 22 September 2018, 05:33:41 UTC | Fix a crash in transpose_times simplification to element times | 22 September 2018, 05:33:41 UTC |
93cc680 | Bowen Bao | 21 September 2018, 17:22:56 UTC | Update reshapeing_test to accomodate python 2.7 | 21 September 2018, 17:24:57 UTC |
d626c2a | Bowen Bao | 21 September 2018, 05:02:57 UTC | Merge branch 'bowbao/gather_backward' | 21 September 2018, 05:02:57 UTC |
da6b0bc | Bowen Bao | 20 September 2018, 21:51:55 UTC | GatherNode backward: add check for no dynamic axis Previously, to resolve issue of gather producing incorrect gradient values, validity mask check was added to ensure we don't count non-valid cells as 0. However, this check is needed only for input that has dynamic axis, i.e. inputs that have MBLayout. | 20 September 2018, 21:54:39 UTC |
b2c28cc | Peyman Manikashani | 20 September 2018, 21:38:02 UTC | updating iteration documentation file | 20 September 2018, 21:39:58 UTC |
0a3eb3b | Bowen Bao | 18 September 2018, 22:00:44 UTC | Update onnx_model_test with tests on cntk pretrained models | 19 September 2018, 18:27:55 UTC |
6f09c39 | Liqun Fu | 18 September 2018, 02:33:30 UTC | Merge branch 'release/2.6' | 18 September 2018, 02:33:30 UTC |
4ed1896 | liqfu | 17 September 2018, 23:10:48 UTC | set public_build to "no"/false | 17 September 2018, 23:10:48 UTC |
1be3b64 | liqfu | 15 September 2018, 01:07:45 UTC | update readme for .net support | 15 September 2018, 01:07:45 UTC |
d355c1c | TJ | 14 September 2018, 19:29:51 UTC | Updated current_iteration with .net support | 14 September 2018, 20:07:24 UTC |
04caa9d | TJ | 14 September 2018, 19:29:51 UTC | Updated current_iteration with .net support | 14 September 2018, 19:29:51 UTC |
7c1b0fa | liqfu | 14 September 2018, 00:01:45 UTC | udpate readme with current iteration | 14 September 2018, 00:01:45 UTC |
da31ba0 | Bowen Bao | 13 September 2018, 20:36:48 UTC | Update current_iteration.md | 13 September 2018, 23:57:44 UTC |
e4d7081 | Sergii Dymchenko | 13 September 2018, 18:18:53 UTC | Update current_iteration.md. | 13 September 2018, 23:57:30 UTC |
4f965aa | liqfu | 13 September 2018, 23:34:30 UTC | update version # in cntk_common.cmake | 13 September 2018, 23:34:30 UTC |
82d350d | liqfu | 13 September 2018, 22:52:47 UTC | bump up version number | 13 September 2018, 22:52:47 UTC |
d264a26 | Bowen Bao | 13 September 2018, 20:36:48 UTC | Update current_iteration.md | 13 September 2018, 20:44:42 UTC |
be28e86 | Sergii Dymchenko | 13 September 2018, 18:18:53 UTC | Update current_iteration.md. | 13 September 2018, 18:18:53 UTC |
deda94b | Bowen Bao | 12 September 2018, 22:08:33 UTC | Support pooling(cpu) where kernel center is on pads. - Previous implementation has the assumption that (0 <= dk < width). This assumption doesn't stand when lo >(kernel - 1) / 2. The updated calculation supports arbitrary lo & hi non-negative integer value. The new calculation has dk in range (0, width + hi + lo]. - Enables onnx backend test {averagepool_2d_pads, maxpool_2d_pads} to pass. | 13 September 2018, 04:37:21 UTC |
62e18f4 | Bowen Bao | 12 September 2018, 22:00:19 UTC | Improve clarity in pads calculation for conv/pool - Refactor function CalcPaddingForSameLowerOrUpperAutoPad in conv/pool import, changing parameter "const Variable& input" to "const NDShape& inputWithBatchAxisShape", to specify the required shape format as [N x C x H x W]. | 13 September 2018, 04:37:21 UTC |
35e3701 | Sergii Dymchenko | 13 September 2018, 02:42:25 UTC | Merge branch 'sedymche/onnx-min-max' | 13 September 2018, 02:42:25 UTC |
d33b7b4 | liqfu | 13 September 2018, 00:24:45 UTC | update iteration plan | 13 September 2018, 00:24:45 UTC |
6a4ec05 | Peyman Manikashani | 13 September 2018, 00:13:46 UTC | Merge branch 'peykash/batchnorm_and_pooling_fixes' | 13 September 2018, 00:13:46 UTC |
61d7dab | Sergii Dymchenko | 07 September 2018, 00:28:02 UTC | Support more than 2 inputs for ONNX Min/Max import. | 12 September 2018, 22:12:14 UTC |
8b48976 | Spandan Tiwari | 12 September 2018, 18:12:10 UTC | Adding CNTK 2.6 release work summary to current_iteration.md | 12 September 2018, 18:12:10 UTC |
b374e14 | Peyman Manikashani | 07 September 2018, 22:41:43 UTC | fixes on Batchnorm and Pooling for v1 pretrained models after removal of sequence axis from input | 12 September 2018, 17:02:52 UTC |
5897265 | Bowen Bao | 07 September 2018, 23:32:51 UTC | small patch on conv/pooling export - when pads are all zero, check if autopad is true. - when pads are all zero, check if ceilOutDim is true, and extra cells are needed. | 11 September 2018, 15:52:22 UTC |
61572e8 | Bowen Bao | 11 September 2018, 15:51:45 UTC | Update onnx_model_test skip list | 11 September 2018, 15:51:45 UTC |
0754b38 | Bowen Bao | 07 September 2018, 00:33:58 UTC | update onnx_model_test with tests from onnx backend test | 09 September 2018, 20:53:26 UTC |
2f52f22 | Bowen Bao | 06 September 2018, 02:22:53 UTC | update conv/convtranspose/pooling import. pad values are explicitly computed based on ONNX spec equations during import in the following cases: - case 1: when auto_pad is SAME_UPPER | SAME_LOWER for convolution, convolution transpose and pooling. - case 2: when output_shape is explicitly set for convolution transpose. note: output_shape in ONNX spec can have the two below format: 1. [X1 * X2 * ... * Xn] 2. [N * O * X1 * X2 * ... * Xn] | 09 September 2018, 20:41:04 UTC |
d877233 | liqfu | 09 September 2018, 15:59:33 UTC | Make broadcast ops compitable between CNTK and ONNX, Enable ONNX export/import for optimizedRNN op, More ONNX support for Sequence ops | 09 September 2018, 15:59:33 UTC |
fcf9f48 | Bowen Bao | 31 August 2018, 18:40:04 UTC | Overhaul conv/convTrans/pooling pads value export - Update exporting of conv/pooling to always export pad values. - Enable correct exporting of multiple pretrained models (ResNet50/ResNet101/ResNet152_ImageNet_Caffe, etc). - Overhaul convtranspose pads exporting - Support conv weight export with omitted out channel axis (LRN). - Add tests in onnx_op_test to cover the above changes | 06 September 2018, 18:46:14 UTC |
e3a1acf | Bowen Bao | 04 September 2018, 23:10:07 UTC | Resolve dependencies and build issues -Temporary add importorskip around import onnx -bump up .yml matplotlib version | 05 September 2018, 22:02:23 UTC |
dc5e482 | Bowen Bao | 30 August 2018, 20:10:12 UTC | fix onnx average pooling export. - this fix solves the issue that ceilOutDim == true will enforce exporting auto_pad as true, even if autoPadding is explicitly set to false. | 31 August 2018, 01:19:04 UTC |
77a8c49 | Bowen Bao | 30 August 2018, 17:49:24 UTC | Temporarily skip onnx_model_test if import onnx fail | 30 August 2018, 17:52:53 UTC |
94a43ed | Liqun Fu | 30 August 2018, 07:40:20 UTC | Merge branch 'liqun/liqun/RNN2.6.Stage' | 30 August 2018, 07:40:20 UTC |
73cd53e | Bowen Bao | 28 August 2018, 23:02:06 UTC | fix nightly issues related to onnx dependencies - Windows OOBE (pip) tests & Linus OOBE tests: skip onnx_model_test. This test requires onnx to be installed. Skip Until we decide to add onnx dependencies to OOBE test environment. | 30 August 2018, 00:19:19 UTC |
18d9f39 | liqfu | 29 August 2018, 23:52:58 UTC | skip dynamic axes wrapper, export onnx test cases, handle output op being Combine op, workaround a specific case bug of ONNX bidirectional broadcast shape inference | 29 August 2018, 23:52:58 UTC |
902f1a4 | Peyman Manikashani | 29 August 2018, 17:08:50 UTC | times export fix | 29 August 2018, 17:09:07 UTC |
b86fe1a | Ke Deng | 28 August 2018, 07:17:54 UTC | Merge branch 'pull/3374' | 28 August 2018, 07:17:54 UTC |
1f2e42e | Spandan Tiwari | 27 August 2018, 07:32:33 UTC | Merge branch 'sptiwari/convtranspose_update7' | 27 August 2018, 07:32:33 UTC |
edc29f8 | Yang Chen | 23 August 2018, 21:28:29 UTC | Packaging newly-added internal header files Recently, we added a couple of new header files into API/Internals. This patch includes them into our pre-built binaries. | 27 August 2018, 04:36:04 UTC |
b3c0fa2 | Spandan Tiwari | 27 August 2018, 04:20:05 UTC | Overhaul ConvTranpose to match ONNX 1.2.2. spec. | 27 August 2018, 04:20:05 UTC |
7dd9638 | Bowen Bao | 24 August 2018, 20:44:28 UTC | squash of the following changes: - fix flatten onnx export. - fix unsqueeze onnx export. - add comments on temporarily skipped tests. - adjust the importing of softmax, logsoftmax and hardmax with blockfunction - such that they could be exported as is back to onnx. - update reshape onnx export to pass mobilenet round trip test. | 26 August 2018, 20:11:44 UTC |
0e20836 | liqfu | 26 August 2018, 15:41:20 UTC | CNTK splice allows broadcast. This case is handled in the change. For noop (identity) ops, its inputs and outputs types shall be set according to upstream ops. ToBatch/ToSequence and Unpack Batch/Sequence ops added during model importing need tp be skipped. Model import need to handle ops with multiple outputs | 26 August 2018, 15:41:20 UTC |
4a6238d | Peyman Manikashani | 23 August 2018, 18:05:56 UTC | reduction all axes export fix | 24 August 2018, 17:02:34 UTC |
d2ff412 | Bowen Bao | 24 August 2018, 04:38:48 UTC | temporarily disable 2 tests on Windows. - This is due to an issue on Windows CI introduced by adding onnx dependencies. These tests are temporarily disabled to not block CI while we investigate. - Disable CNTKv2Python/Tutorial/205 - Disable CNTKv2Python/Keras | 24 August 2018, 08:27:45 UTC |
88d8882 | Phoebe Ma (Beyondsoft Corporation) | 24 August 2018, 05:56:36 UTC | Fix issue#3228 and #3363 which found when build with MSVC+ permissive- | 24 August 2018, 05:56:36 UTC |
a26e542 | Bowen Bao | 23 August 2018, 22:21:34 UTC | update conftest.py to resolve doctest issue. - newer version numpy has a different print format for arrays and scalars that would potentially break the doctests. | 24 August 2018, 04:28:06 UTC |
28ada96 | Bowen Bao | 23 August 2018, 06:09:53 UTC | update python doctest handling newer version numpy print format. | 23 August 2018, 06:21:13 UTC |
56ef694 | Bowen Bao | 21 August 2018, 18:34:49 UTC | add onnx_model_test.py | 23 August 2018, 06:21:13 UTC |
c0ff1da | Bowen Bao | 21 August 2018, 18:35:08 UTC | fix gemm, pooling export to onnx. | 23 August 2018, 06:19:18 UTC |
4244320 | Peyman Manikashani | 21 August 2018, 20:59:03 UTC | Adding support for exporting CNTK TimesTranspose | 22 August 2018, 18:33:55 UTC |
3d809bf | Yang Chen | 15 August 2018, 22:28:58 UTC | Added several internal API header files In case other projects may use these header files, we added them into API/Internals. * ComputationGraphAlgorithms.h was moved from Source/ComputationNetworkLib * PrimitiveOpType.h and EvaluatorWrapper.h were moved from Source/CNTKv2Library * PrimitiveFunctionAttribute.h was extracted from PrimitiveFunction.h. It contains a new class PrimitiveFunctionAttribute which is the collection of all attribute names for PrimitiveFunction. This change actually had a subtle side-effect. We had a global static variable s_stateAttributes that depended on PrimitiveFunction::AttributeNameRngSeed and PrimitiveFunction::AttributeNameRngOffset. After we moved those static attribute-variables into another translation unit, s_stateAttributes can be initialized with empty wstring, because PrimitiveFunctionAttribute::AttributeNameRngSeed PrimitiveFunctionAttribute::AttributeNameRngSeedOffset were initialized after s_stateAttributes. Note that the initialization order of global static variables is not well-defined cross translation units. To fix the issue, we also moved s_stateAttributes into PrimitiveFunctionAttribute class, and renamed it to s_rngStateAttributes. I think it's reasonable to consider s_rngStateAttributes to be part of the PrimitiveFunctionAttribute class. * PrimitiveFunction.h was moved from Source/CNTKv2Library | 22 August 2018, 17:47:18 UTC |
8a244e6 | Bowen Bao | 21 August 2018, 21:22:59 UTC | fix onnx_op_test.py reduce ops. - remove bracket around data to match latest verification. | 21 August 2018, 21:23:28 UTC |