Commit graph

7863 commits

Author SHA1 Message Date
Ke Zhang
139abda393
convinteger implementation based on gemmlowp (#294)
* update

* cmake change

* rename

* update

* update

* add cmake

* fix build warnings.

* fix comments

* update cmake to avoid run gemmlowp tests

* update cmake

* update

* fix build break

* update

* fix comments

* fix test failure

* add one more test case with padding.

* fix conv implementation of mkldnn and cuda to use updated computekernelshape function.

* fix linux ci build break
2019-01-15 14:39:50 -08:00
Hector Li
835b511fa8 cuda fix to unblock the tf model tests (#333)
* Check the pads attribute on Conv, and auto fallback to CPU if it's not symmetric padding

* Insert copy nodes after all graph transformer. It causes some issue if do the cast transformer before memory copy transformer.
2019-01-15 14:05:47 -08:00
Changming Sun
7977871740 Split build pipeline 2019-01-15 12:30:59 -08:00
jignparm
7e3923b9b3
Fix for non-wide characters in strings for linux - for c#-native interop (#326)
* Fix for non-wide characters in strings for linux - for c#-native interop

* update some unit tests

* added unicode and utf-8 encoding explicitly for file names
2019-01-15 01:41:32 -08:00
Hector Li
779123cf55
Upsample opset 9 cuda implementation (#330) 2019-01-14 23:04:46 -08:00
Raymond Yang
0efc48a11a
Install dotnet sdk on linux ci (#320)
* Try install dotnet sdk on linux ci

* Fix install script

* Add configurable os version in docker build script

* Avoid use ARG in docker
2019-01-14 17:51:45 -08:00
Edward Chen
677918cd9a Added generation of C# project properties file containing actual build directory. 2019-01-14 16:02:13 -08:00
Shah Asaduzzaman (ASAD)
c955cd8278 changed csharp runtime folder name to win from win10 2019-01-14 16:01:57 -08:00
Changming Sun
25c1e68988 Fix: roi_pool operator implementation error about FLT_MIN 2019-01-14 15:21:37 -08:00
Changming Sun
ef5679949a
Fix a c# build issue when mkldnn is not enabled (#321) 2019-01-14 14:22:02 -08:00
jignparm
3b83f062fc remove delayload from mkldnn (#276) 2019-01-14 14:13:28 -08:00
Changming Sun
260639c327 Add missing EXCLUDE_FROM_ALL keyword to nsync submodule 2019-01-11 16:34:55 -08:00
Faith Xu
2d067ec65c Update with link for C# GPU Nuget package 2019-01-11 15:39:18 -08:00
Yufeng Li
4735bb1ccb
Add onnx protobuf format that supports large model (#313)
* Add onnx protobuf format that supports large model

* Add optional for DataLocation
2019-01-11 10:23:18 -08:00
Sreekanth Yalachigere
05b9440fce mkldnn:Conv weight optimization (#256)
* mkldnn:Conv weight optimization

* weight optimization: review changes

* lock_guard and mutex for thread safe

* mutex added to provider

* lock to ReOrder done only once

* removed #ifndef mkldnn_hpp

* keep re-ordered mem buffer in scope

* applied clang format

* review updates: map to unordered map

* conv_mutex to mutex_
2019-01-11 08:48:20 -08:00
Xavier Dupré
8c40313e28
Update documentation to reflect the latest changes (#311)
- removes markdown output
- rename intro into index
- uses skl2onnx anywhere possible instead of onnxmltools
2019-01-11 12:41:42 +01:00
Du Li
7641ee9a2b suppress a warning. 2019-01-10 19:28:18 -08:00
Bowen Bao
d22429c5b2 Update compare_mlvalue for tests (#290)
* values should be considered matched if both of them are inf, or both
of them are nan.
2019-01-10 18:13:46 -08:00
Changming Sun
02962ce9d8
Update ABI.md (#299)
* Update ABI.md
2019-01-10 16:34:42 -08:00
Yufeng Li
02852a0881
Remove OperatorParser tool (#279) 2019-01-10 16:14:37 -08:00
Pranav Sharma
223773d278
Implement ROI Align for object detection. (#308)
* Implement ROI Align for object detection.

* Fix Mac build

* Fix Mac build
2019-01-10 11:34:55 -08:00
Changming Sun
6b3044ddd3 Update AddingExecutionProvider.md 2019-01-10 11:11:54 -08:00
Randy
fa0ea9a273
implement dynamic slice cuda (#286)
* implement dynamic slice cuda

* add template parameter

* add delaration

* init base class

* exclude case from cuda

* use cuda mapped type

* separate function implementation

* add cpy logic

* refactor

* add type check

* use InputMemoryType

* merge functions
2019-01-10 09:42:18 -08:00
Ryan Hill
98a92547bf
Ryanunderhill/c api 8 (#297)
* Make OrtAllocator not be reference counted

* Make the allocator interface more type safe

* Fix build break

* Build break fix

* Build break fix

* Mistake in previous build fix.

* Fix review comments + build break

* Missed the export symbols

* C specific error, need 'struct' keyword in one case.

* Function calling OrtReleaseObject instead of OrtReleaseEnv
2019-01-10 02:06:29 -08:00
Changming Sun
751eb60819 Fix a build issue in onnxruntime_unittests.cmake (#307) 2019-01-09 22:53:53 -08:00
KeDengMS
d342147255
Add test for truncated sequence inference (#99)
* Add test for truncated sequence inference with scan model

* Address CR

* Update to Scan opset 9
2019-01-09 22:33:35 -08:00
edgchen1
34bcc92554 Added test data URL and checksum arguments to build.py. (#302)
* Added test data arguments to build.py, modified win-ci-pipeline build.

* Updated CI builds to use template tasks, added test data args, removed AZURE_BLOB_KEY uses.

* Fixed up set test data step template.
2019-01-09 22:33:14 -08:00
Yufeng Li
55433abecb Enable build_shared_lib if build_csharp 2019-01-09 22:32:17 -08:00
Changming Sun
84be07114b Disable nsync test 2019-01-09 18:56:12 -08:00
Yulong Wang
22b5bfdf35 update label for build badges (#301) 2019-01-09 16:54:41 -08:00
Changming Sun
e318c7317b update 2019-01-09 15:49:27 -08:00
Changming Sun
8cfe8d33a3 Add nsync (#292)
* Add nsync

* nsync2

* nsync3

* fix build

* update comments

* fix build option
2019-01-09 10:40:55 -08:00
Hector Li
57421504b1
Add tensorflow license (#296) 2019-01-08 12:43:27 -08:00
Pranav Sharma
31bbb4598e
Enable tvm in CI builds. (#285)
* Enable tvm in CI builds

* Fix tvm dll path issue
2019-01-07 19:37:06 -08:00
Ke Zhang
386b60e35a
update gemmlowp which carries a windows build break fix. (#291) 2019-01-07 18:24:46 -08:00
Yang Chen
4929ddde43 update tvm submodule (#287) 2019-01-07 14:32:58 -08:00
Scott McKay
059cfdc13a
Opset 9 Scan implementation (#274)
* Templatize Scan as step 1

* Pre-thunderstorm save

* Initial v8 and v9 implementations.
Need to add transpose to v9 and unit tests.

* Make Transpose operator implementation re-usable by Scan.
Add transpose logic to Scan.

* Rework a bunch of things. First Scan 9 unit test passes

* Add more tests.

Need to add axis validation and handling of negative values.

* Convert remaining Scan 8 tests to also work for Scan 9 if applicable.
Add invalid input tests for new Scan 9 attributes.

* Add transpose unit test.
Some cleanups.

* Cleanups

* Check number of direction entries for outputs at kernel instantiation.
2019-01-08 08:11:46 +10:00
Changming Sun
5e113661a9 Build system upgrades (#281)
* update

* runas normal user
2019-01-07 13:15:24 -08:00
Tang, Cheng
85ec13f58d
fix tvm break (#282) 2019-01-07 10:55:24 -08:00
Raymond Yang
ec2cf59baa
Enable building python37 packages (#283) 2019-01-05 18:41:40 -08:00
Pranav Sharma
de383d93be
Fix inconsistency in enum names in the C API (#277)
* Fix inconsistency in enum names in the C API

* fix build
2019-01-04 16:41:15 -08:00
Ke Zhang
251d4bbb23
adjust the quantization ops' definitions (#278) 2019-01-04 16:05:29 -08:00
shahasad
8a7b6761de
changed nativeLib name from onnxruntime.dll to onnxruntime, for cross platform (#275) 2019-01-03 19:52:21 -08:00
Ke Zhang
75934af896
have Im2ColNd support all types and allow customized padding value. (#273)
* have Im2ColNd support all types and allow customized padding value.

* only specialize the template in order NCHW.

* fix build break.

* fix build break
2019-01-03 19:24:06 -08:00
Dmitri Smirnov
058803086d
Implement N-gram (#180)
* Implement N-gram
  Do not load unnecessary pool n-grams. Add String typed tests.
  Set output size to the mav ngram_index value plus 1.

* Address security warnings and some review comments.

* Fix build issues, rework sampling to try all n-gram sizes at a given offset.

* Rework the loop so all n should be tried at a given offset
  and we do not add the same items all over again such as
  b,c and next we try b,c,d but we no longer add b,c again.

* Compute hash incrementally so we do not re-hash elements that were
  already there when we add more elements to n-gram.

* Address review comments.
  TODO: Remove all attribute.

* Remove all attribute, adjust tests. Correct docs.

* Address more review comments.

* Create Type And Shape inference function.

* Address review comments. Implement batch mode per new spec.

* Correct switch bracing in OutputResult and re-test.

* Fix shape error message within TypeAndShapeInferenceFunction.
2019-01-03 16:21:03 -08:00
ashku-ms
8fba324678
Hyperbolic inv ops (#272)
Implement Inverse for hyberbolic ops

Eigen will add support for asinh, acosh and atanh in the upcoming release. But until then for completeness of opset9 we have std based implementation.
2019-01-03 15:04:55 -08:00
Tang, Cheng
d0fa974976
interface change to code-generated kernels (#192)
* merge function compile interface

* fix build error

* fix linux build break

* fix static cast issue; fix clang style

* fix argument change

* use alignment allocation;fix comments in pr

* fix linux break

* apply clang format

* rename according to comments in pr

* rename according to pr comments;remove useless file

* remove the need_compile flag

* avoid passing whole session state
2019-01-02 17:18:08 -08:00
xkszltl
5b6f1823af Use CUDA libraries found by CMake. (#190)
* Link cudart and use CMake predefined variables for cublas.

* Patch for old cmake.

* Remove static cudnn.
Fix file name for cuBLAS DLL delay loading.
2019-01-02 16:51:42 -08:00
Randy
fc76076e29
Rashuai/dynamic slice refactored (#264)
* define dynamic slice

* remove obsolete

* add test cases

* remove disabled cases

* rename test cases

* fix comments

* format code

* fix comments

* fix compile err

* fix typo

* removed duplicated delaration

* add enforced checks

* add enforced checks

* add extra processing on negative axis

* fix typo
2019-01-02 16:39:41 -08:00
Yuan Yu
bd2ace7619 Tweak linker flags to minimize the binary size. (#270) 2019-01-02 14:50:34 -08:00