Commit graph

185 commits

Author SHA1 Message Date
Changming Sun
1ed72251d0
A tiny change to onnxruntime::Model (#158)
* Remove unnecessary const

* update
2018-12-12 15:33:20 -08:00
Ke Zhang
9baaf5e956
Kezhan/change edge api to use index (#138)
* remove input edges while removing a node.

* put comments on how to remove a node without using resolve.

* fix test failure

* fix test failure.

* fix format

* try to guess and fix mac ci failure.

* change add remove edge api.

* update IR

* update comments

* fix comments.
2018-12-12 14:30:55 -08:00
Changming Sun
2e79597531
Revert "mkldnn Relu/Sum/BatchNormalization kernels" (#147)
* Revert "mkldnn activations.relu (#86)"

This reverts commit 6d80253502.

* Revert "MKLDNN Sum and Batch Normalizaton kernels (#115)"

This reverts commit bdd5e58546.
2018-12-12 14:28:52 -08:00
Du Li
2230e5b431
Conv+Activation fusion for CPU (#105)
* Add conv+activation fusion.

* Adding tests

* Adding activation LeakyRelu.

* Refactoring the code to use a fusedConv custom op instead of changing
the original conv op at runtime.

* fix build issue.

* fix build issue.

* In order to reduce binary size:
1. reuse onnx shape inference for conv
2. remove most doc.

* Accomodating PR comments.

* Accomodating PR comments

* Remove unused variables
2018-12-12 11:37:09 -08:00
Pranav Sharma
7f0e5269bd
Removing from broken tests since this failure cannot be reproduced any more. (#154) 2018-12-11 20:35:04 -08:00
Yuan Yu
f189b76f9a Some small edits and renaming. (#153) 2018-12-11 16:36:39 -08:00
Changming Sun
0ae9354db1
Delete unused files in onnx_test_runner (#149) 2018-12-11 14:41:34 -08:00
Prasanth Pulavarthi
396e95bc4b
Reduce whitespace at top of readme for easier reading (#148)
* Reduce whitespace at top of readme for easier reading

* Update README.md
2018-12-11 13:52:47 -08:00
Ke Zhang
60c6b9dfd0
Kezhan/separate op def from kernels (#143)
* separate ops from kernels so that folks don't need to depend on kernels when doing export

* resolve conflicts.

* fix comments and warnings in linux

* fix build break in linux

* fix build break.

* fix the wrong dependency that framework needs providres.
2018-12-11 11:57:27 -08:00
Raymond Yang
e2a90ab747 Add unique identifier in function subgraph node (#129) 2018-12-11 11:53:20 -08:00
Changming Sun
1beeccb89d Enable data cache (#144) 2018-12-10 19:15:03 -08:00
Yufeng Li
8510ccdfe7 Update performance_runner.cc (#142) 2018-12-10 17:33:20 -08:00
Ashwini Khade
b054646ddd
Askhade/implement erf (#137)
* erf implementation for op9

* enable erf node tests + review comment fixes

* update CMAKE flag

* plus erf to execution provider
2018-12-10 17:26:01 -08:00
Pranav Sharma
7d79bfef71
Move isnan out of contrib_ops and add float16 support for it as per the spec. (#141)
* Move isnan out of contrib_ops and add float16 support for it as per the spec.

* Remove isnan from list of broken tests
2018-12-10 16:12:56 -08:00
Changming Sun
9bf78e1f3e
Update BUILD.md (#140) 2018-12-10 15:56:39 -08:00
Changming Sun
7fb44b9d9e
Add missing #include (#139) 2018-12-10 13:57:07 -08:00
Ashwini Khade
c5a691d67d
update onnx to latest commit (#132)
* update onnx to latest commit

* fixes

* Add not implemented operator tests to whitelist

* fix test name
2018-12-10 11:17:52 -08:00
Hector Li
69a5ff3300
Upsample opset 9 support for CPU provider (#135)
* Upsample opset 9 support for CPU provider

* remove template from UpsampleBase

* resolve build error on Linux

* Fix build warning
2018-12-10 11:15:29 -08:00
Zhang Lei
d71376010b
Implement Dynamic Range Operator with shape inference
Implement dynamic range operator in custom ops, with shape inference.
2018-12-10 10:12:53 -08:00
Sreekanth Yalachigere
bdd5e58546 MKLDNN Sum and Batch Normalizaton kernels (#115)
* mkldnn activations.relu

* updates after PR review 1

* mkldnn Sum and Batch Normalization Kernels

* BN,Sum PR Review changes

* Sum PR review 2

* Sum PR review 3

* PR Review 4: const and const refs
2018-12-07 22:58:59 -08:00
jignparm
6bfb195184
Jignparm/csharp test difftypes (#126)
* Added tests for C# types

* added multithread c# test

* Added pretrained tests

* Added function to load data from PB file

* added exec task for protoc
2018-12-08 01:34:08 +00:00
Dmitri Smirnov
50ed36290c
Use windows specific C-runtime based solution for lowercasing/uppercasing (#131)
* Use windows specific C-runtime based solution for lowercasing/uppercasing.
  This worksaround cpp runtime bug with std::locale construction.

* Address review comments.
2018-12-07 16:24:52 -08:00
jywu-msft
ed5abc8b94
mkldnn conv 1d, 3d support. (#130) 2018-12-07 14:48:19 -08:00
shahasad
a09a3d3aa5
Add build targets and props file for legacy csharp project support (#127)
* added the x64 condition in the targets file

* need the reference node for managed dll

* address review comments

* fix the dll path

* added back Link
2018-12-07 12:15:11 -08:00
Siva
19c630b87f [TVM] Bugfix. (#117) 2018-12-07 10:31:16 -08:00
Yufeng Li
23e122f304
Run backward gru in main thread for bidirection (#118) 2018-12-07 10:24:57 -08:00
Ashwini Khade
6169f3c619
First Draft EyeLike CPU OP9 (#121)
* First Draft EyeLike CPU OP9

* Plus review comments

* moving to eigen

* naming fixes
2018-12-07 10:15:50 -08:00
shahasad
5cec70df4b added the redirected URL to nuget package icon (#125) 2018-12-06 22:21:44 -08:00
Pranav Sharma
c5cbdd5a55
Miscellaneous fixes (#123) 2018-12-06 22:21:04 -08:00
shahasad
a53d617b7a
added the icon and logo resources (#124) 2018-12-06 21:22:11 -08:00
jignparm
9b9cb32f88
Jignparm/csharp test difftypes (#122)
* Added tests for C# types

* added multithread c# test
2018-12-07 05:17:31 +00:00
Ke Zhang
830c341c19
remove graph editor since it's not designed as expected to restrict graph access for rewrite rule. (#119) 2018-12-06 14:10:51 -08:00
Weixing Zhang
194d0a98d1 Add domain and version check for ops fusion. (#108)
* Add domain and version check for ops fusion.

* Create a helper function for OpType, domain and version check.

* Put domain as a parameter.
2018-12-06 11:24:37 -08:00
Tracy Sharpe
3c7c1068e7
refactor threading (#110) 2018-12-06 09:20:32 -08:00
Sreekanth Yalachigere
6d80253502 mkldnn activations.relu (#86)
* mkldnn activations.relu

* updates after PR review 1

* modified license

* Added MIT License to header
2018-12-05 19:33:56 -08:00
Dmitri Smirnov
c52636e187
Implement Tokenizer op (#31)
* Implement separator tokenizer with TST.
  TODO: Clarify what to do if the output is empty and no start/end text
  markers required. Also see if the current search algo is acceptable.

* Add utf8 util test

* For empty output produce [C] -> [C][0], [N][C] -> [N][C][0]

* Augument TST search with match conflict resolution in favor of
  earlier specified pattern matches.

* Address MAcOS build error.

* Adjust error message

* Address review comments.

* Remove nested loops.

* Remove 3rd party utf8 validation code.

* Address review comments part I.

* Move padding outside start/end markers.
  Split unit tests for invidividual test cases.

* Fix a common prefix bug reported by Xavier.
2018-12-05 17:52:04 -08:00
Hector Li
a68f5ccfd9
Upgrade gpu build to CUDA 10 + cudnn 7.3 (#112)
* Upgrade gpu build to CUDA 10 + cudnn 7.3

* update the yaml file for python package building

* switch to the cuda9.1 docker file if the CUDA_VER is cuda9.1-cudnn7.1
2018-12-05 17:49:16 -08:00
Hector Li
4801e67104
Hecli/cuda10 upgrade (#111)
* Add build step to remove the cuda msbuildcutomization file after build, otherwise, the cuda high version could impact the lower version build

* update vs path

* update the path
2018-12-05 14:58:40 -08:00
Ryan Hill
eea7618d53
Ryanunderhill/c api (#85)
* Unify all C API header files into a single file.
Remove Ptr suffix and use '*' syntax directly.
2018-12-05 14:03:45 -08:00
Tang, Cheng
dd1b16dd58
add brainslice change in common folder (#109) 2018-12-05 13:48:51 -08:00
Hector Li
3dcf344f09
switch build agents to the CUDA 10 pool (#106) 2018-12-05 12:48:02 -08:00
Xavier Dupré
0573952499 Update the documentation, run all examples during the generation of the documentation (replace #89) (#103)
* Minor update in the documentation

* Run examples during the generation of the documentation.
2018-12-05 10:12:25 -08:00
Klein Hu
eeb862ab1a Update Dockerfile for ARM build and the build instruction (#93)
* Update Dockerfile for ARM build and build description

* Update BUILD.md for ARM build

* Fix typo

* Kezhan/renaming graph_base.h to graph.h (#95)

* rename graph.h to graph_viewer.h

* rename graph_base.h to graph.h

* 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.

* Split the CI pipelines (#94)

* Add the split build badges (#96)

* Split the CI pipelines

* Update build bdge display table in README

* Revert Softmax optimizations using openmp. (#97)

Revert "Simpler unused parameter in #if defined() switch."
This reverts commit 6b00e6bb4d.

Revert "Better opemmp parallel group count calculation in Softmax parallel running."
This reverts commit c530064ebe.

Revert "Optimize softmax cpu by parallel using openmp."
This reverts commit e7bdfa00db.

* Delete logo as per marketing requirements (#98)

* Update Dockerfile for ARM build and build description

* Update BUILD.md for ARM build

* Fix typo
2018-12-04 23:06:36 -08:00
R. G. Esteves
c8ae91251e Eliminate the confusing double negative (#92)
* Eliminate the confusing double negative

I was having trouble parsing the caveat NOTE, proposing wording changes that I think reflect the meaning and avoid the confusion.

* Eliminate double negative without further explanation on role of this file.

Incorporate @pranavsharma feedback.
2018-12-04 23:06:09 -08:00
Pranav Sharma
358be0653c
Delete logo as per marketing requirements (#98) 2018-12-04 18:04:59 -08:00
Zhang Lei
996d6ea4cd
Revert Softmax optimizations using openmp. (#97)
Revert "Simpler unused parameter in #if defined() switch."
This reverts commit 6b00e6bb4d.

Revert "Better opemmp parallel group count calculation in Softmax parallel running."
This reverts commit c530064ebe.

Revert "Optimize softmax cpu by parallel using openmp."
This reverts commit e7bdfa00db.
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