saymrwulf/onnxruntime
1
0
Fork 0
mirror of https://github.com/saymrwulf/onnxruntime.git synced 2026-06-24 02:47:54 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
11997 commits 2856 branches 57 tags 1.3 GiB
Commit graph

11997 commits

This branch
This branch
All branches
Author SHA1 Message Date
jywu-msft
7763f62a09
fix ngraph build (#1062) 
* fix ngraph due to incompatibility with PR1019

* temp disable onnx backend test test_operator_repeat_dim_overflow_cpu due to failure in ngraph ep

* disable tiny_yolov2 for ngraph

* temporarily disable operator_repeat_dim_overflow for ngraph due to graph resolve failure
 2019-05-18 09:21:59 -07:00
Dmitri Smirnov
58d9f6a145 Enable greater/less tests. (#1061) 2019-05-17 21:15:12 -07:00
Yufeng Li
f90bd234ff
fix warning in scatter op (#1054) 2019-05-17 17:15:44 -07:00
Doris Xin
fd5eb3b351 Remove dropout nodes during inference (#1018) 
* initial version. there's a bug

* allow nodes with multiple outputs to be removed if only one output is used by downstream ops

* move node output checks into their own methods

* add test data file

* address comments by @kkaranasos and @skottmckay

* address more comments by @kkaranasos

* a comment to clarify the position of the mask output
 2019-05-17 15:05:36 -07:00
Changming Sun
2663b9c443
Remove unnecessary casts from OrtValue to MLValue(#1051) 2019-05-17 07:52:59 -07:00
Ke Zhang
a7039601c4
setting input/output for partitioned function subgraph (#1019) 
* setting input/output for partitioned function subgraph

* fixed TensorRT related failure

* fix failures for tvm and ngraph

* update
 2019-05-17 11:19:28 +08:00
jignparm
11069765dc
Fix C-API sample (causing internal build failure (#1047) 2019-05-16 15:49:39 -07:00
Changming Sun
99556b111d
Make MemPatternPlanner on/off switchable in model weight loading (#989) 2019-05-16 14:39:09 -07:00
Changming Sun
7bce377113
Fix LTO build failure on ubuntu (#1048) 
* Fix LTO build failure on ubuntu
 2019-05-16 14:06:38 -07:00
Darío Hereñú
2324f65a22 Fixed typo on paragraph 14 (#997) 2019-05-16 11:30:36 -07:00
Changming Sun
e4225f8234
Combine OrtValue and MLValue into one type (#1043) 
* Combine OrtValue and MLValue into one type
 2019-05-16 10:22:49 -07:00
Scott McKay
6c65640708
Update SessionState::GetKernel to just call find() instead of count() and find(). (#1046) 
Update SessionState::GetInputNodeInfo to just call find() instead of count() and at().
 2019-05-16 10:02:01 -07:00
Tracy Sharpe
2e19b14e4e
Fuse elementwise add/multiply into convolution (#1028) 
* fix elementwise fusions

* update comments

* test case
 2019-05-15 20:54:21 -07:00
jignparm
d14e65a224
Finer control over when Python tests are run (#1023) 
* Finer control over when Python tests are run

* add --build_wheel to linux pipeline, instead of run_build.sh

* add --build_wheel to all ci configurations

* update per review comments
 2019-05-15 18:30:18 -07:00
jywu-msft
303f3c1278
drop GIL before Run to enable multiple python threads to execute in parallel (#1042) 
* drop GIL before Run to allow multiple python threads to execute in parallel.

* revert setup.py change
 2019-05-15 18:11:14 -07:00
Changming Sun
d3c0cd24fc
Change the cpu allocator info used in OptimizerExecutionFrame class (#1036) 
* Change the cpu allocator info used in OptimizerExecutionFrame class
 2019-05-15 16:42:51 -07:00
Changming Sun
5062be612e
update (#1041) 2019-05-15 16:41:41 -07:00
jignparm
fc17a8bc08 Fix MIN and MAX aggregation functions (#1022) 2019-05-15 10:19:25 -07:00
Maik Riechert
875c4c2f3f restore ninja compatibility 
The ninja generator requires all generated libraries to be declared.
This was missing in get_boost.cmake.
 2019-05-15 10:18:52 -07:00
Changming Sun
ec2a8b2386 delete test/onnx/simple_thread_pool.h 2019-05-15 10:00:59 -07:00
Changming Sun
e42099480e
Clean up code (#1033) 
Some trivial changes for making gcc compiler happy. Won't impact runtime behavior.
 2019-05-15 10:00:39 -07:00
Changming Sun
46bf6fd1ef
Replace fabs with std::fabs (#1035) 
fabs only accept double, there it needs float.
Similar thing for log.
 2019-05-15 10:00:12 -07:00
Scott McKay
9029c496b0
Currently download_test_data in build.py supports only the standalone version of 7-Zip by looking for 7za.exe. Add support for the standard version which has 7z.exe. (#1030) 2019-05-16 02:38:38 +10:00
Scott McKay
3ac4826268
Don't force the install of a specific numpy version unless explicitly requested to when running build.py. Validate that just the minimum version required is available. (#1031) 2019-05-16 02:38:18 +10:00
Tracy Sharpe
a63598ac5d
remove old name tables (#1029) 
Remove old hash tables from inference session.
 2019-05-14 21:46:13 -07:00
Yufeng Li
0f6b3ea575 Improve the performance of Scatter (#991) 
* Improve the performance of scatter
 2019-05-14 13:35:15 -07:00
tmccrmck
cb25fd4d8a Updates ARMv7 Dockerfile to use newer CMake (#1017) 2019-05-14 13:33:44 -07:00
Maik Riechert
f5dfde33d0 generalise strerror_r preprocessor switch (#1026) 
same approach as in https://reviews.llvm.org/D25071
 2019-05-14 13:11:32 -07:00
Xavier Dupré
d32d2fbfb7
Fix svm runtime (replaces #91, #33) (#101) 
* Fix svm runtime
 2019-05-14 22:04:18 +02:00
Klein Hu
406770c484 Enable ONNX Runtime Server Model Tests (#1002) 
* Enable model tests for ORT Server in build script

* Nightly build pipeline definition

* Force prune docker images

* Clean up containers before prune images
 2019-05-13 23:08:51 -07:00
Klein Hu
f7e57a3d16 Prune containers and images (#1003) 
* `docker image prune -f`to clean up unused images

* Prune exited containers
 2019-05-13 16:57:17 -07:00
Yufeng Li
9330a42fae
Optimize upsample (#1008) 
Extract the index and scale computation out of the loop to improve the perf
 2019-05-13 16:30:02 -07:00
Dmitri Smirnov
30f87c7526
Update build instructions based on recent experience. (#1011) 
Clarify instructions with respect to Linux/Windows differences and add additional info.
 2019-05-13 14:38:49 -07:00
Ryan Hill
3408494407
More C++ API improvements and conversions (#998) 
* More C++ API improvements and conversions
* Mark more constructors as explicit
* Fix CSharp function name changes
* Change more test cases to use C++ API
 2019-05-13 13:56:54 -07:00
Hariharan Seshadri
c69dff7928
Implement contrib kernels for Pad (changed interface) and Unique (new ONNX op) (#1006) 
* Intial commit

* Rename DynamicPad to Pad

* More changes

* Add Unique operator

* Revert accidental check-in

* Fix CUDA Pad to align with changes

* More changes

* Fix more CUDA pad source files

* More fixes

* More changes

* More changes

* Avoid vector copy

* Update vector validation logic

* Fix build failures

* Fix build

* Fix build failure

* Fix tensorrt build
 2019-05-13 13:10:18 -07:00
Hariharan Seshadri
49a2928a59
Support string type for split op (#1014) 
* Initial commit

* Support string type in split

* More changes

* More changes

* Minor nit

* Revert nit fix
 2019-05-13 11:21:55 -07:00
KeDengMS
31cbb5d33d
Add support for int32_t in CPU provider's Where op (#1009) 
This is required for running BiDAF model. https://github.com/onnx/models/pull/157
 2019-05-13 10:08:27 -07:00
Konstantinos Karanasos
4e5fdd3167 Removing one of the generate tranform tests (#1010) 2019-05-12 18:01:08 -07:00
Mika Fischer
c0acb8b6c3 Allow loading model from in-memory byte-array (#718) 2019-05-11 05:58:50 +08:00
tmccrmck
af7090c25f Fix server's executor memory leak and add tests (#984) 
* Fixes memory leaks in executor and adds executor tests

* Remove logging test from executor_test.cc

* Uses RAII to allocate and freee MLValue buffers

* Change private member buffer to buffer_
 2019-05-09 22:51:57 -07:00
Hariharan Seshadri
3fb24f6409
Improve error messaging based on raised issues (#988) 
* Initial commit

* Changes to help improve end user facing messages

* Resolve comments

* Revert "Resolve comments"

This reverts commit 2bec877fea025250f2d07b0bd0c1d54b6d0c8aae.

* Resolve comments V2

* Fix nit

* Resolve comments

* Clean up condition to return error message

* Resolve nit

* More changes
 2019-05-09 11:43:41 -07:00
Du Li
b08dbd37cb
Adding Min cuda kernel (#992) 2019-05-09 10:06:30 -07:00
Pranav Sharma
2de1f43a40 Move EP docs to the docs folder. (#996) 
* Move docs to the docs folder

* Move docs to the docs folder

* Move docs to the docs/execution_providers folder.
 2019-05-09 09:46:18 -07:00
jignparm
a463f765fc
Minor fix to pass x86 native build (#994) 2019-05-09 00:22:45 -07:00
Harry Yang
2512b0ebeb Adding the onnxruntime Dockerfile and instructions (#968) 
* Adding the onnxruntime Dockerfile and instructions

* Adding the onnxruntime server Dockerfile with build and instructions

* Specify more details in instruction

* Update dockerfiles/README.md

Co-Authored-By: tianchijushi <huaimingyang@hotmail.com>

* Update Dockerfile.server

Specify the onnxruntime server build branch during git clone
 2019-05-08 18:24:16 -07:00
Changming Sun
9a4128efac
Enable model test in Mac pipeline (#990) 2019-05-08 14:43:13 -07:00
Klein Hu
c2b412f7be Update the ONNX Runtime Server CI pipeline setup (#986) 
* Update the ORT-SRV ci pipeline setup

* Update pip package installation for server tests

* Install requests package in build setup

* Check if python dependencies exists before install
 2019-05-08 11:37:39 -07:00
Vinitra Swamy
47171a076b
Addressing #956, deprecating old CPU and GPU quickstart dockerfiles (#987) 2019-05-08 10:40:01 -07:00
Changming Sun
9263932861
perf test: fix an index out-of-bound bug (#982) 
1. Fix an index out-of-bound bug
2. Fix a memory leak in the TF test runner
 2019-05-07 21:40:33 -07:00
Konstantinos Karanasos
403fb703a0
Avoid using transformers that rely on contrib ops in tests (#983) 2019-05-07 17:23:04 -07:00