saymrwulf/onnxruntime
1
0
Fork 0
mirror of https://github.com/saymrwulf/onnxruntime.git synced 2026-05-24 22:17:32 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
4969 commits 2856 branches 57 tags 1.3 GiB
Commit graph

4969 commits

This branch
This branch
All branches
Author SHA1 Message Date
Gao, Chun
4dd724ef1a
Enable WebAssembly SIMD build (#7839) 
Add a build switch "--enable_wasm_simd" to enable
WebAssembly SIMD build
 2021-05-28 16:29:58 -07:00
Ryan Hill
5a63904aa9
Remove some templated versions of functions that are no longer needed (#7868) 
* Switch to non template version of function
 2021-05-28 13:22:45 -07:00
Maajid khan
e41e042de6
[OpenVINO-EP] Adding OpenVINO-EP samples to Msft Repo (#7826) 
* Added ONNX_OV_EP samples

->Added cpp, python and csharp samples
using OpenVINO Execution Provider.

Signed-off-by: MaajidKhan <n.maajidkhan@gmail.com>
 2021-05-28 08:35:41 -07:00
Edward Chen
ab4b5055c7
[Objective-C API] Fixes from package testing and clean up (#7866) 2021-05-27 19:36:50 -07:00
Edward Chen
35b49b64c7
Fix regex to detect Objective-C/C++ (.m/.mm) files. (#7870) 2021-05-27 19:35:59 -07:00
Hariharan Seshadri
0255c83dc4
Clean up CPU kernel definition for opset 13 Pad (#7867) 2021-05-27 19:32:36 -07:00
Tianlei Wu
71b05f74a2
fix duplicated node name (#7865) 2021-05-27 17:16:17 -07:00
Guoyu Wang
1f4421fe70
Include ORT C/C++ API headers in the ORT Mobile AAR package (#7858) 
* Add header files of ort c/c++ api to aar package

* Move header file selection to cmake based on EP choice
 2021-05-27 17:07:48 -07:00
Scott McKay
63df683040
Fix path used in check for cudnn library (#7786) 
* There are separate paths for CUDA and CUDNN as they are not guaranteed to be in the same location on a Windows machine. Use the CUDNN path when looking for the CUDNN library.

* Refine check
 2021-05-28 09:32:13 +10:00
baijumeswani
ddf4aaaae1
Resolve issue with wrapped ORTModule load_state_dict (#7847) 
* Encapsulate children modules inside a ModuleAccessor object to prevent erroneuos iteration over children while loading the state dictionary

* Add named_models, models, apply methods, change ModuleAccessor to ModuleMetadata and modify unit tests

* Change ModuleMetadata module getter logic, raise NotImplementedError for add_modules

* Add comment explaining why overriding _load_from_state_dict method is needed
 2021-05-27 16:11:37 -07:00
Chen Fu
8140e3fde5
Make requantize a qgemm post processor (#7850) 
Description:
Change requantize interface so it can be processed block by block. This enable as to make requantize to be a post processor of QGEMM.

Motivation and Context

Previous changes show we improve performance by parallelize batch gemm. Unfortunately we could not parallelize the batch gemm in quantize_linear_matmul due to the requantize operation at the end of each gemm. By changing requantize to be a qgemm post processor, we now can parallelize the batch operation.

Co-authored-by: Chen Fu <fuchen@microsoft.com>
 2021-05-27 15:05:04 -07:00
Yulong Wang
ccdedf1b2e
[js] update documents (#7852) 
* [js] update documents

* escape double quotes

* update operators.md

* resolve comments
 2021-05-27 14:51:57 -07:00
Hariharan Seshadri
7380219717
Fix bug in Transpose CUDA kernel (#7329) 2021-05-27 14:01:08 -07:00
Hariharan Seshadri
883923a40a
Support bool type for Pad CPU (#7856) 
* Initial commit

* update

* nit
 2021-05-27 13:48:45 -07:00
Edward Chen
13622bae91
Add Apple log sink. (#7820) 
Add a log sink for Apple platforms. This version uses NSLog().
 2021-05-27 10:03:02 -07:00
Edward Chen
45a7352622
Update Mac CI builds to use macOS-10.15 image, Xcode 12.4. (#7437) 
Update Mac CI builds to use macOS-10.15 image, Xcode 12.4.
 2021-05-27 09:39:34 -07:00
Tixxx
2a3851cd75
fixed bugs in packed mode and enable pack mode tests in ci (#7848) 
* fixed bugs in packed mode and enable pack mode tests in ci

* removed unnecessary space

* pr comments

* pr comments

* disable an average pool test

* try disabling another avg pool

* disable more avg pool tests

* disable maxpool tests
 2021-05-27 07:56:58 -07:00
liqunfu
bed6e87cbd
add environment variable to control default training package's local version (#7849) 2021-05-26 22:44:20 -07:00
Edward Chen
fa093d8e45
[Objective-C API] Add ORTSession methods to get input, overridable initializer, and output names. (#7837) 2021-05-26 19:54:55 -07:00
Yufeng Li
94bb09bf47
fix topo sort in quant tool (#7833) 
* fix topo sort in quant tool

* add unit test and make the topo sort stable
 2021-05-26 17:53:35 -07:00
Sherlock
fc472a04be
Relax tol for Conv1D fp16 test (#7844) 
* Relax tol for Conv1D fp16 test

Co-authored-by: Sherlock Huang <bahuang@OrtTrainingDev3.af05slrtruoetgaxwwjv5nsq5e.px.internal.cloudapp.net>
 2021-05-26 17:04:35 -07:00
Siva Popuri
c08bb4eee3
Update docs/ONNX_Runtime_Server_Usage.md (#7818) 
Making it clear in the documentation to proactively inform users.
 2021-05-26 16:17:20 -07:00
Guoyu Wang
afca89dce6
fix boost download url (#7843) 2021-05-26 16:08:57 -07:00
Yulong Wang
331f20428c
[js/web] only apply max thread number when it's omitted (#7834) 2021-05-26 15:46:50 -07:00
Thiago Crepaldi
c5ea5907c0
Fix permission error for ORTModule lock file (#7814) 2021-05-26 14:18:25 -07:00
Hariharan Seshadri
c487824a31
Fix bug in Einsum implementation (#7822) 2021-05-26 10:27:19 -07:00
Dmitri Smirnov
d1f0251e39
Python bindings fix ups in preparation to Sparse Tensor introduction (#7817) 
* Fix up constness in pybindings
  Fix up return argument treatments.
  Specifically, for all functions that return pointers or references
  to the members of other pybind registered classes, we want not to copy
  them, but internally bump up a reference to the hosting class so they do not
  disappear before the reference to the returned members is re-claimed.
  This policy is applied by default to def_property and def_readwrite but not to def_readonly
  and other def methods.
  See https://pybind11-jagerman.readthedocs.io/en/stable/advanced.html#return-value-policies
  https://pybind11.readthedocs.io/en/stable/advanced/functions.html#return-value-policies
  Move OrtValue binding to a separate file
  Move IOBinding into separate file.
 2021-05-26 09:47:41 -07:00
Scott McKay
57782b3463
Add supported operators/types documentation for the ORT Mobile package (#7807) 
* Add ability to generate documentation for the ORT Mobile package using the build configuration as input.
 2021-05-26 15:57:40 +10:00
harshithapv
4fe59c8b29
delete model_copy to save memory allocated in forward call (#7832) 
* delete model copy

* add flag

* address comments

* address flag comment

Co-authored-by: root <root@OrtTrainingDev0.af05slrtruoetgaxwwjv5nsq5e.px.internal.cloudapp.net>
 2021-05-25 22:22:13 -07:00
George Wu
1c6b6f696e
fixes for cuda centos/manylinux (#7830) 
* fixes for cuda centos/manylinux

* remove providers_shared.so dep processing.
 2021-05-25 19:38:59 -07:00
Ryan Hill
f78af4fc8c
Use RTLD_GLOBAL for onnxrutime_providers_shared on unix (#7831) 
* Use RTLD_GLOBAL for onnxrutime_providers_shared on unix
 2021-05-25 19:03:24 -07:00
Tixxx
ea1a4f8fb5
[JS]support running super resolution model using ortweb (#7677) 
* migrated changes to support running super resolution model using ortweb

* reverted benchmarking tool related changes which will be in a separate pr

* added kernel tests to op and node tests

* minor change to the order of variables

* added one more unit test for packed matmul
 2021-05-25 17:43:43 -07:00
Pranav Sharma
6ca1ee7733
Fix rpath issue with pybind. (#7829) 
* Fix rpath issue with pybind

* Address PR comment
 2021-05-25 17:36:15 -07:00
Jesse Benson
29c68888af Update BERT convergence baseline. 2021-05-25 17:11:46 -07:00
Jesse Benson
3d12e957a7 Workaround for miopenReduceTensor() behavior difference in ROCm 4.2 2021-05-25 17:11:46 -07:00
Zhang Lei
f49a4b6329
Decrease lock contention in qlstm by memory allocation. (#7815) 
* Decrease lock contention in qlstm caused by memory allocation.
 2021-05-25 17:08:42 -07:00
ytaous
ff655175ff
Eliminate no op node - add 0 (#7798) 
* eliminate add 0

* typo

* rank check

* fix build

Co-authored-by: Ethan Tao <ettao@OrtTrainingDev4.af05slrtruoetgaxwwjv5nsq5e.px.internal.cloudapp.net>
 2021-05-25 13:01:34 -07:00
Ryan Hill
9241d76396
Remove unnecessary cuda libraries refernced in cmake (#7824) 2021-05-25 10:01:15 -07:00
Changming Sun
93c8e29782
Improve code coverage report (#7770) 2021-05-25 08:26:01 -07:00
Hariharan Seshadri
cdd2129fed
Fix CUDA Pad kernel registrations (#7813) 2021-05-24 21:11:19 -07:00
Xavier Dupré
a41255c280
Fix performance regression in Reduce operators introduced by PR #7206 (#7719) 
* improves ArgMin implementation
* update  parallelization cost
* choose former implementation for KRK case, when K=1
* improves unit test
 2021-05-25 01:41:12 +02:00
Pranav Prakash
f487f6be25
Add int8/int32 Relu for Opset 14 (#7536) 
* Add int8 Relu for Opset 14

* update kernel def hashes and exclude TensorRT for int8 relu

Co-authored-by: Hari Vallabhaneni <harivall@berkeley.edu>
 2021-05-24 13:19:14 -07:00
Guoyu Wang
98007f0be6
Fix typo in the ios packaging script (#7802) 2021-05-24 11:57:13 -07:00
Suffian Khan
02c78a8aa8
test migration to rocm4.2 (#7800) 2021-05-24 11:48:44 -07:00
Ashwini Khade
da5ab325ee
register batchnorm for opset14 cuda provider (#7806) 2021-05-24 11:02:45 -07:00
Sheil Kumar
bd5067a2ff
Cannot upgrade SDK version because winml_lib_telemetry pulls in SDK cppwinrt version (#7795) 
* fix telemetry includes

* add dependencies

Co-authored-by: Sheil Kumar <sheilk@microsoft.com>
 2021-05-24 08:00:24 -07:00
baijumeswani
13a129054f
Prevent unnecessary re-initialization of the graph when model has unused parameters (#7799) 2021-05-22 20:52:26 -07:00
Scott McKay
c4f515d380
- Fix training cmake file so it builds if --cmake_extra_defines onnxruntime_BUILD_UNIT_TESTS=OFF is specified. (#7789) 
- Fix check on cudart_versions when building on Windows to handle None being returned
 2021-05-23 09:53:15 +10:00
stevenlix
de4c221712
set max profile before opt profile (#7801) 2021-05-22 09:45:13 -07:00
Guoyu Wang
ae14cedd63
Fix c_api warning (#7803) 2021-05-22 01:23:39 -07:00