Zhang Lei
996d6ea4cd
Revert Softmax optimizations using openmp. ( #97 )
...
Revert "Simpler unused parameter in #if defined() switch."
This reverts commit 6b00e6bb4dc530064ebee7bdfa00db
2018-12-04 16:21:27 -08:00
Raymond Yang
7f0e947f96
Add the split build badges ( #96 )
...
* Split the CI pipelines
* Update build bdge display table in README
2018-12-04 15:01:44 -08:00
Raymond Yang
3b08c6665a
Split the CI pipelines ( #94 )
2018-12-04 13:51:35 -08:00
Dmitri Smirnov
fbb23a9ed0
Implement StringNormalizer ( #69 )
...
* Imlpement StringNormalizer
Add mixed language tests, test case insentive path.
* Create a locale on the fly. Default locale does not seem to create well.
* Add CI language-pack-en to make default locale available.
Catch and translate locale creation exception to make the message
meaningful.
* Make sure locales are configured on Ubuntu.
2018-12-04 13:47:08 -08:00
Ke Zhang
005f9dca96
Kezhan/renaming graph_base.h to graph.h ( #95 )
...
* rename graph.h to graph_viewer.h
* rename graph_base.h to graph.h
2018-12-04 13:25:39 -08:00
Ke Zhang
a78acb2d2c
rename graph.h to graph_viewer.h ( #84 )
2018-12-04 08:41:03 -08:00
Changming Sun
980e8aa270
Update BUILD.md ( #90 )
2018-12-04 08:40:27 -08:00
Brian Martin
744ec28c88
(de)quantize_linear optimizations ( #76 )
2018-12-04 08:39:27 -08:00
Pranav Sharma
37385ec029
Add Dockerfile and page for ARM builds. ( #83 )
2018-12-03 19:13:59 -08:00
Raymond Yang
1e59b6f1c2
Minor fixes to CI definitions ( #80 )
...
* Align win gpu ci parameters with vsts definition
* Fix linux CI script to avoid docker name conflict
2018-12-03 17:35:23 -08:00
jywu-msft
47a6992e1b
update mkldnn to 0.17.1 and address assumptions related to tensor padding that come with new mkldnn version. ( #79 )
2018-12-03 17:07:44 -08:00
Hector Li
900e69ceae
User lower case while comparing the activation method as it's not clear in the spec. ( #77 )
...
Fix the bug for RNN/GRU/LSTM auto fall back, the supported activation should cover bidirectional mode.
Update the test to remove the flag to enable gpu test since we can auto fallback to cpu
2018-12-03 14:21:23 -08:00
shahasad
f1c66a4aae
Updated CSharp package description ( #75 )
...
* Updated CSharp package description
* capitalize Runtime
2018-12-03 13:54:58 -08:00
Hector Li
1c9d0b2729
Add missing types for Slice op ( #74 )
2018-12-03 13:54:48 -08:00
Weixing Zhang
aa549cd194
Support fusing 3D Conv with Add/Mul. ( #23 )
...
* Support fusing 3D Conv with Add/Mul.
With this PR, the subgraph 3D Conve->Add->Mul in Resnet3D can be fused into one 3D Conv.
* Updated it based on feedback.
* Updated it based on review feedback.
* Change the implementation of scale_by_axis rather than Mul.
* Refactor the code to make the compiler optimize it easily.
2018-12-03 13:11:04 -08:00
Pranav Sharma
b7cc611563
Minor documentation changes ( #78 )
2018-12-03 12:55:29 -08:00
Faith Xu
07253fd647
Faxu patch 1 ( #63 )
...
* Delete Roadmap.md
* Update README.md
* Update README.md
* Update README.md
2018-11-30 22:10:26 -08:00
Pranav Sharma
21dad1780b
Add logo and link to C# pkg. ( #73 )
2018-11-30 19:39:30 -08:00
Tracy Sharpe
28bb96be2a
use std::copy to avoid gsl::span overhead ( #71 )
2018-11-30 18:44:41 -08:00
Raymond Yang
fd0d7c5fc0
Add the mac python packaging script ( #72 )
...
* Add pipeline for building python wheels for Windows/Linux CPU and GPU
* try enable mkldnn
* remove mklml
* Update python packaging configuration
* Try macos pywheel packaging
* Try removing mkldnn from mac build
* Use conda in mac agents
* Change to build release only
* Add the mac wheel packaging list to the packaging yaml
* Add mkldnn into mac wheels
2018-11-30 17:31:15 -08:00
Hector Li
31780cae32
Add NonMaxSupression op to contribution ops ( #60 )
...
* Add NonMaxSupression op to contribution ops
* fix the warning
* Update schema, add more test case
* fix build error on Linux
* update according to review comments
* fix typo
* add shape inference
* revert a minor change
2018-11-30 17:15:13 -08:00
Pranav Sharma
6624dd2778
Rel 0.1.5 ( #70 )
...
* updated nuget package metadata for MS compliance (#66 )
* fixed metadata element -- use PackageProjectUrl instead of ProjectUrl (#67 )
* Change version to 0.1.5
2018-11-30 16:23:47 -08:00
Tracy Sharpe
21044846ee
use #pragma parallel for properly ( #68 )
2018-11-30 14:11:49 -08:00
shahasad
c55663860e
fixed metadata element -- use PackageProjectUrl instead of ProjectUrl ( #67 )
2018-11-30 11:01:47 -08:00
Zhang Lei
b534f9fa5f
Square and Cube optimization for Pow<float>() operator. ( #30 )
2018-11-30 10:16:22 -08:00
shahasad
bcc8f621ea
updated nuget package metadata for MS compliance ( #66 )
2018-11-30 09:35:20 -08:00
Scott McKay
edbe19d22f
Include output name with mismatch error message so it's clearer where the failure comes from. ( #65 )
2018-11-30 20:34:35 +10:00
Ke Zhang
9b1bc06c34
Merge pull request #64 from Microsoft/kezhan/remove_const_cast
...
remove const cast
2018-11-29 19:16:05 -08:00
linkerzhang
e9de9abc96
update
2018-11-29 18:36:05 -08:00
linkerzhang
fbd7b9b8f0
Merge branch 'master' of https://github.com/Microsoft/onnxruntime into kezhan/remove_const_cast
2018-11-29 18:27:21 -08:00
linkerzhang
56811d10af
remove const cast in transformer_memcpy
2018-11-29 18:27:06 -08:00
linkerzhang
e00c956254
update the comments.
2018-11-29 17:54:30 -08:00
Pranav Sharma
39ebccbc8b
Fix sample example documentation for python pkg ( #61 )
2018-11-29 17:50:12 -08:00
linkerzhang
afe3fb5c1d
Merge branch 'master' of https://github.com/Microsoft/onnxruntime into kezhan/remove_const_cast
2018-11-29 17:46:03 -08:00
linkerzhang
35f94cac27
remove const cast from conv related fusion.
2018-11-29 17:45:44 -08:00
Xavier Dupré
1ea32a097a
Update pythin bindings for options ( #14 )
2018-11-29 17:03:09 -08:00
Pranav Sharma
728c3c078e
Update issue templates ( #62 )
2018-11-29 15:37:40 -08:00
Zhang Lei
cd1042c94f
Merge pull request #36 from Microsoft/zhalei/softmax_optimize
...
Optimize softmax cpu by parallel using openmp.
2018-11-29 13:52:26 -08:00
Ke Zhang
fafc48bf94
Merge pull request #59 from Microsoft/kezhan/qlinearmatmul
...
add QLinearMatMul
2018-11-29 11:36:52 -08:00
linkerzhang
91c860bec4
fix description
2018-11-29 11:17:22 -08:00
linkerzhang
34aa545b64
add QLinearMatMul
2018-11-29 10:58:20 -08:00
Ke Zhang
e8db06ed44
Merge pull request #32 from Microsoft/kezhan/op_spec_fix
...
fix output type specification for ConvInteger
2018-11-29 08:48:05 -08:00
shahasad
7780fd6cd9
fix bug -- failing on non float tensor output types ( #58 )
...
* fixed the missing implementation of output tensor construction based on type checking
* cleanup
* cleanup
2018-11-29 08:15:18 -08:00
Scott McKay
8980cbbfad
Merge pull request #52 from Microsoft/scmckay/UpdateCudaInfoInBuildMd
...
Update CUDA build help and tweak build setup
2018-11-29 18:06:57 +10:00
Scott McKay
a4bcb1121b
Merge pull request #57 from Microsoft/scmckay/FixInferenceSessionInputValidationHandlingOfOptionalInputs
...
Support overriding initializers via feed inputs
2018-11-29 18:05:43 +10:00
Scott McKay
97dc949e3f
Fix some more C# unit test issues.
...
Ignore bin and obj directories under /csharp
2018-11-29 17:38:51 +10:00
Scott McKay
f3fb1d5ff8
Merge remote-tracking branch 'origin/master' into scmckay/FixInferenceSessionInputValidationHandlingOfOptionalInputs
2018-11-29 16:35:05 +10:00
Scott McKay
2486336355
Update error messages C# tests check for.
2018-11-29 16:31:10 +10:00
Scott McKay
d263d6734e
Adjust test. valid input names aren't deterministically ordered in error message.
2018-11-29 15:04:53 +10:00
Scott McKay
4dd1e50aa7
Merge remote-tracking branch 'origin/master' into scmckay/UpdateCudaInfoInBuildMd
2018-11-29 14:53:11 +10:00
