saymrwulf/onnxruntime
1
0
Fork 0
mirror of https://github.com/saymrwulf/onnxruntime.git synced 2026-06-04 23:59:56 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
1388 commits 2856 branches 57 tags 1.3 GiB
Commit graph

1388 commits

This branch
This branch
All branches
Author SHA1 Message Date
Changming Sun
c24d7a8a0a
Update eigen to the latest version (#1910) 2019-10-11 10:44:19 -07:00
Scott McKay
bdfff800ea Move access to intra-op threadpool into OpKernelContext. (#2091) 2019-10-11 10:36:20 -07:00
Changming Sun
368bdfd936
Update README.md (#2070) 
Update the vcredist package link

Note: Visual C++ 2015, 2017 and 2019 all share the same redistributable files.
 2019-10-11 10:06:50 -07:00
Hector Li
3b335c933f
Fix issue that TRT not work for device other than device id 0 
Fix issue that TRT not work for device other than device id 0. Because the allocation planner need to get the default allocator to allocate memory for graph input data. (#2094)
 2019-10-11 09:22:25 -07:00
Scott McKay
ffb94fd170
Fix bug with delayed allocation of If and Scan outputs. (#2024) 
* Fix bug with delayed allocation of If and Scan outputs.
If the subgraph is producing output on a non-CPU device the delayed allocation was incorrectly providing a CPU allocated tensor.
Check for the required location, and update 'fetches' instead if a device copy is needed.
The utils::ExecuteGraph logic will handle the device copy in this case.
 2019-10-11 19:49:21 +10:00
Yang Chen
ca1b88c069
Added support to infer Pad11 (#2085) 
* Added support to infer Pad11

* address CR
 2019-10-10 23:18:49 -07:00
shahasad
8803f6fff4
C# end to end test fix, and make end to end tests mandatory (#2079) 2019-10-10 19:23:43 -07:00
Changming Sun
a314402097
Downgrade python gpu package to CUDA 10.0 (#2086) 2019-10-10 18:31:24 -07:00
Dmitri Smirnov
af9dbb70f2
Introduce a separate check and conditional for AVX512BW build (#2083) 
Separate checks for AVX512f and AVX512BW
  Make AVX512BW cmake instructions nested within AVX512F support.
 2019-10-10 16:14:00 -07:00
Hariharan Seshadri
2ba705ed99
Handle nodes with subgraphs in ORT function handling implementation (#2053) 
* Initial commit

* Update

* Update

* Nits

* More updates

* to be reverted

* Update

* Update

* More changes

* Updates

* Update Function

* Nits

* Fix build break

* Comment
 2019-10-10 16:07:42 -07:00
Pranav Sharma
2d4d0abd36
Add support for output seq(tensor) in python binding and test framework. Implement SequenceConstruct, SequenceEmpty, SequenceInsert and SequenceErase ops. (#2040) 
Add support for output seq(tensor) in python binding and test framework. Implement SequenceConstruct, SequenceEmpty, SequenceInsert and SequenceErase ops. (#2040)
 2019-10-10 15:58:49 -07:00
Scott McKay
ddbc2086e4 Add support for opset 11 Clip in optimizers. (#2059) 2019-10-10 10:47:29 -07:00
Yulong Wang
a41c71cbf2
check and fix CUDA kernel launch errors in several OPs (#2047) 2019-10-10 23:47:00 +08:00
baowenlei
b4a98aab78
change MatMulInteger/MatMulInteger16 fallback option (#2064) 
* change MatMulInteger/MatMulInteger16 fallback option when no initializer exist

* add AVX option

* fix condition for old machines
 2019-10-09 22:03:21 -07:00
Hariharan Seshadri
d186c19c45
Add opset-11 TopK CPU kernel (#1912) 
* initial commit

* Update

* Update top_k.cc

* PR comments

* Add more tests

* Update

* Add another test case

* Update

* Resolve conflicts

* Update

* Nits

* Nits

* Nits

* Pick sorted content using 2 different approaches

* Update to logic

* PR comments

* PR feedback

* Update

* Fix build

* Fix build

* Update
 2019-10-09 19:09:30 -07:00
Colin Versteeg
8fda6593fe Update failing tests (#2038) 
* Fix failing tests from when they were not enabled

* split into two

* fix failing test
 2019-10-09 15:17:21 -07:00
Tracy Sharpe
57e0099425
MLAS: Implement U8S8 GEMV kernels (#2069) 
This implements an optimization for U8S8 MlasGemm when M=1, aka GEMV.
 2019-10-09 11:54:16 -07:00
Changming Sun
eee9c55030
C++11 fix for memcpy_transformer_test.cc (#2061) 2019-10-09 10:52:10 -07:00
Changming Sun
cefae93305
Add a test case for linearregressor (#1962) 2019-10-09 10:17:08 -07:00
Changming Sun
ccaf692ff2
Run auditwheel for manylinux1 (#2063) 2019-10-09 09:23:00 -07:00
Dmitri Smirnov
cae571c713 Add a test for AVX512 compilation before compiling 512 asm (#2055) 2019-10-08 21:18:04 -07:00
Changming Sun
af8fe0f980
Replace make_unique in cuda_utils.cu (#2052) 2019-10-08 18:32:08 -07:00
Scott McKay
db0dd09ded
Cleanup some aspects of the Initializer class used by optimizers (#2005) 
* Move check on data type outside of the Initializer class as it's specific to Conv processing.
Use references for arguments that can't be null.
 2019-10-09 10:37:44 +10:00
Changming Sun
a00ca56ae1
Remove gcc from manylinux1 docker image (#2048) 2019-10-08 13:49:15 -07:00
baowenlei
b82de794d5
Weba/update nuphar doc (#2026) 
* update nuphar xp doc

* address comments

* address CR

* update doc
 2019-10-08 12:41:25 -07:00
RandySheriffH
f501b6e234
pack pyop in nightly build (#2018) 
* pack pyop in nightly build

* correct logic

* add comment

* exclude debug build

* add dependency

* reset postbuild rule

* remove dep
 2019-10-08 12:02:45 -07:00
Changming Sun
e9bed8b23b
Change python packaging pipeline to use manylinux1 (#2035) 
1. Change the python packaing pipeline to use manylinux1
2. Temporarily disable model test in the python pipeline.
 2019-10-08 10:03:54 -07:00
Changming Sun
3053af812c
Fix a crash in deep_cpu_gru_op_test.cc (#2028) 2019-10-08 10:03:07 -07:00
Zhang Lei
71b389322e Implement cuda scatter op. (#1991) 
* Implement cuda scatter op.
Disable Invalid Index of Scatter op only for cuda provider.

* Fix some pipeline's type narrow warning as error.
 2019-10-08 09:53:33 -07:00
Yang Chen
a94c9bd88d throw exception using dmlc::LogMessageFatal (#2033) 
* throw exception using dmlc::LogMessageFatal

On windows, ORT_THROW couldn't be caught if the exception was thrown from
a jitted functions. Let's call dmlc::LogMessageFatal instead.

* address CR

use LOG(FATAL)
 2019-10-08 09:31:35 -07:00
Yang Chen
19b0d0af87
Enabled bool input type for Equal for op_ver 11 (#2034) 
This change enabled bool type for Equal-11's inputs
 2019-10-08 01:50:37 -07:00
Yang Chen
203c2f5b59
updated reduce_ops for op_ver 11 (#2039) 
After enabling op_ver 11 for reduce ops, we need to check axes to
make sure it's not empty.
 2019-10-08 01:05:05 -07:00
Pranav Sharma
f13b66768a
Fix build for gcc 4.8.5. (#2036) 2019-10-08 00:50:53 -07:00
shahasad
b70fc34fae
Fix C# end to end tests in NuGet pipeline, failing for missing test data file 2019-10-07 20:14:20 -07:00
shahasad
b0feaef9de
Update the C# pretrained model test to include opset9 and 10 models (#2003) 2019-10-07 19:14:34 -07:00
George Wu
0bd807f3b3
trt provider status return cleanup (#2032) 
* status and code cleanup.

* revert change. seems like a bug in TRT causes intermittent failure return?
 2019-10-07 18:34:48 -07:00
Tianlei Wu
b2c1937523
Add EmbedLayerNormalization and SkipLayerNormalization ops for bert optimization (#2012) 
* Add Embed Layer Normalization and Skip Layer Normalization ops for bert optimization.

* add float16 test for skiplayernorm

* Add test for EmbedLayerNormalization op

* fix cpu build error

* fix build warning

* update HasCudaEnvironment function

* handle cuda error
 2019-10-07 17:29:43 -07:00
Changming Sun
8f7657fa32
Ignore some gcc warnings (#1996) 2019-10-07 16:32:34 -07:00
Pranav Sharma
ea60469af5
Support seq(tensor), implement 2 sequence ops that use the new type. (#1983) 
* Mention OrtCreateSessionFromArray in C API doc

* fix seq of tensors

* changes on 9/30

* All tests passing

* Add SequenceAt op

* Fix shared_lib non_tensor_types test

* Address some PR comments

* Address PR comments

* Add support in python bindings to accept seq(tensor)

* Change data type from vector<Tensor> to TensorSeq

* Change data type from vector<Tensor> to TensorSeq

* Added some documentation

* Added missing test model

* Fix Linux build

* Fix Mac build

* Fix Mac build
 2019-10-07 15:35:09 -07:00
Hector Li
00e24ae4fe
refactor Cuda Ops Sum, Max, Min, remove dup code (#1946) 
refactor Cuda Ops Sum, Max, Min, remove dup code
 2019-10-07 13:17:49 -07:00
Tianlei Wu
7b39f5090c
Add Attention op for multi-head self attention in BERT (#1984) 
* Add Attention op for multi head self attention in BERT

* Add test cases

* Move op from kOnnxDomain to kMSDomain.
Limit test to run by CUDA provider only.

* fix test

* Add float16 test

* fix cpu build error

* handle cuda error

* get last cuda error when failed
 2019-10-07 12:22:54 -07:00
Yang Chen
7d2f0c79bd Bumped up to op_ver 11 for a bunch of Nuphar Ops (#2025) 
This change enabled op_ver 11 for a dozen of Nuphar Ops
 2019-10-07 10:34:05 -07:00
Changming Sun
3c26ae5b6d
ThreadPool fix for roialign and CropAndResize (#2020) 2019-10-06 22:43:59 -07:00
Pranav Sharma
4cdb95e436
Resort to sequential execution if the inter op thread pool ptr is nullptr; (#2023) 2019-10-06 16:08:41 -07:00
stevenlix
544e53e24e Update TensorRT to version 6.0.1.5 (#1966) 
* remove onnx-tensorrt submodule

* add new onnx-tensorrt submodule (experiment) for trt6

* update engine build for trt6

* update compile and compute for tensorrt6.0

* Update tensorrt_execution_provider.cc

* Update tensorrt_execution_provider.cc

* Update tensorrt_execution_provider.cc

* Update tensorrt_execution_provider.cc

* switch to onnx-tensorrt master for TensorRT6'

* Update tensorrt_execution_provider.cc

* Handle dynamic batch size and add memcpy in TensorRT EP

* update test cases

* Update tensorrt_execution_provider.cc

* update onnx-tensorrt submodule

* Update Dockerfile.ubuntu_tensorrt

* Update Dockerfile.ubuntu_tensorrt

* Update run_dockerbuild.sh

* Update run_dockerbuild.sh

* Update install_ubuntu.sh

* Update concat_op_test.cc

* Update tensorrt_execution_provider.cc

* Upgrade TensorRT to version 6.0.1.5

* Update onnxruntime_providers.cmake

* Update CMakeLists.txt

* Update reduction_ops_test.cc

* Update install_ubuntu.sh

* Update Dockerfile.ubuntu_tensorrt

* Update Dockerfile.tensorrt

* Update BUILD.md

* Update run_dockerbuild.sh

* Update install_ubuntu.sh

* Update onnxruntime_providers.cmake

* Update install_ubuntu.sh

* Update install_ubuntu.sh

* Update gemm_test.cc

* Update gather_op_test.cc

* Update CMakeLists.txt

* Removed submodule

* update onnx-tensorrt submodule

* Add Ubuntu18.04 build option

* Add Ubuntu18.04 build option

* Add Ubuntu18.04 build option

* Add Ubuntu18.04 build option

* Remove redundency

* Fix issue that it does not add memcopy node correctly if some nodes fall back to CUDA EP.
e.g. after partition, there's TRT_Node -> Cuda_node (with CPU memory expected), we still need to add memcpy node between them.

* update for Trt Windows build

* Update onnxruntime_providers.cmake

* Disable opset11 tests on TensorRT

* Update pad_test.cc

* Update build.py

* update scripts for ubuntu18.04

* Disable warning for Windows build
 2019-10-06 10:40:53 -07:00
baowenlei
4bb6385dca
Weba/merge ngemm (#2021) 
* save status: add tiling layout; add avx512 skylake cpuid info

* unit tests and matmul integer model passed on skylake, need to verify model

* save commit before update master

* fix check

* address comments
 2019-10-05 12:09:22 -07:00
Xavier Dupré
0b5aac0a2e fix python setup (#2022) 2019-10-05 09:46:41 -07:00
Yang Chen
e8285a7996
Added GatherElements to Nuphar (#2016) 
* Added GatherElements to Nuphar

This change added GatherElements (op_ver 11) to the Nuphar provider.

* address CR feedback

* create a utilify function for accessing index safely

* address more CR

* SafeIndex -> ClampIndex
 2019-10-04 23:53:02 -07:00
Colin Versteeg
1ba76c5f74 add support for empty version and score route (#1995) 2019-10-04 22:53:11 -07:00
Changming Sun
a9e04a29b3
Ignore a test: ParallelExecutor.StatusPropagation (#2019) 2019-10-04 22:51:47 -07:00