saymrwulf/onnxruntime
1
0
Fork 0
mirror of https://github.com/saymrwulf/onnxruntime.git synced 2026-06-26 03:00:54 +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
Jeff
e89dd92387 Flush and trim resources in DML EP in new OnSessionInitializationEnd method 2020-04-14 21:40:53 -07:00
Changming Sun
b63349c8d6
Fix custom op test failure (#3525) 2020-04-14 20:36:42 -07:00
Adam Pocock
bc9a199b16 Renaming deviceNum to deviceId. 2020-04-14 20:35:03 -07:00
Adam Pocock
e9dc8954ac Adding support for ACL and DML to the Java API. 2020-04-14 20:35:03 -07:00
pengwa
2c7c45076b
MaxBatchSize E2E Test (#3454) 
* max batch size e2e test

*update test data snapshot
 2020-04-15 09:50:44 +08:00
Changming Sun
a2feb29b0d
Fix build break (#3528) 
Ignore some known test failures
Install ONNX package before running Windows CI builds
 2020-04-14 18:07:56 -07:00
Jeff
1950391570 Specific DML version latest and fix compile error 2020-04-14 16:38:23 -07:00
Negin Raoof
e303f458e4
Add int64 input type for ReduceProd (#3507) 
* Add int64 input type

* Fix for cuda

* Fix linking

* Cuda

* Fixed missing registration

* Fix registeration for opsets 1-11

* Adding reduce_matrix_rows for int64

* Update reduction_functions.cu

* Revert cuda
 2020-04-14 15:09:28 -07:00
edgchen1
4fa88a0a23
Remove cast to OpKernelContextInternal to get threadpool and directly use OpKernelContext. (#3523) 2020-04-14 14:30:26 -07:00
Ori Levari
f564569a80
Adapter Model and Environment tests (#3469) 
*Adapter Model and Environment tests
*winml test macro clean up and extension
 2020-04-14 13:36:31 -07:00
Tiago Koji Castro Shibata
560f4c5b16
Make GPUTEST macro consistent among TAEF/googletest (#3518) 2020-04-14 10:55:16 -07:00
Du Li
621b3ac03a
FFT contrib ops (#3381) 
* add custom op skeleton

* Adding Rfft, Irfft kernels.

* Fix a few errors:
1. make kernel stateless to avoid race condition
2. reclaim cufft plan

* Adding MLFloat16 support

* Adding fp16 support for fft ops.

* Adding cufft plan cache.

* adding a util func

* adding copyright info.

* Accommodating PR comments.
 2020-04-14 10:12:04 -07:00
Tixxx
06b63975c0
Fix fp16 type mismatch when graph output is an fp32-only node (#3411) 
* verify output node before changing its type in mixed precision mode
 2020-04-14 09:35:19 -07:00
Yufeng Li
baa86f181f
Handle the case that initializers are in graph input (#3449) 
warn that initializers are in graph input
provide a tool to move initializer out of graph input
Motivation and Context
ONNX model from IR_VERSION 4 only treats initializers that appear in graph input as non-constant. This may fail some of the graph optimizations, like const folding, operator fusion and etc. Warn the case and provide a tool.
 2020-04-14 09:06:04 -07:00
David Brownell
006c5be1b1
Optionally produce a python wheel that includes featurizers (#3491) 2020-04-14 09:00:13 -07:00
Changming Sun
040c28ff39 Remove dead code from HandleNegativeAxis 2020-04-14 01:01:15 -07:00
edgchen1
ba7225f986
Update Graph SetInputs and SetOutputs for training (#3446) 
Fix training modification of Graph SetInputs() and SetOutputs(). Originally there were distinct code paths in Graph based on whether the graph was loaded from a GraphProto or created from scratch. The training modifications made that distinction a bit ambiguous - i.e., even though the Graph is loaded from a GraphProto for training, sometimes we rely on the other code path, e.g., to deduce the graph inputs after modifying it. Consequently, there was some odd behavior when using SetInputs(). For correctness, this change separates the cases where the graph is loaded from a GraphProto and where it is created from scratch.
 2020-04-13 19:10:44 -07:00
Colin Jermain
06db89cf13 Using logic for finding README.rst to find requirements.txt 2020-04-13 18:59:44 -07:00
Colin Jermain
43d9f9190e Removing unused six package 2020-04-13 18:59:44 -07:00
Colin Jermain
c2c3102aba Tying install_requires to requirements.txt 2020-04-13 18:59:44 -07:00
Ye Wang
66a79d2c9f
fix (#3512) 2020-04-13 18:30:58 -07:00
Dmitri Smirnov
efd9b92482
Handle Scalars in TernaryOps and Where. (#3509) 
Handle Scalars in TernaryOps and Where.
 2020-04-13 16:24:35 -07:00
Ye Wang
cbe30f3e19
update FeaturizersLibrary (#3511) 2020-04-13 15:47:51 -07:00
Tracy Sharpe
5aab2671f8
Fix crash in DequantizeLinear with scalar tensor (#3508) 2020-04-13 14:52:52 -07:00
Ye Wang
438353abcd
Fix TruncatedSVDFeaturizer's test failure and re-enable it's kernel test (#3458) 
* checkin

* fix linux & macos build

* fix test

* revert the changes for a single-aimed PR

* fix
 2020-04-13 13:59:38 -07:00
Jeff
67443c0e17 Fix winml test compilation errors 2020-04-13 13:06:19 -07:00
M. Zeeshan Siddiqui
5d99f179b9
Merge pull request #3486 from microsoft/sedymche/merge_master_ort_training 
Merge from master into ort_training
 2020-04-13 10:55:36 -07:00
Jeff Bloomfield
d1384ba27e Merge branch 'user/jeffbloo/MergeGithubMasterToDmlDev' of https://microsoft.visualstudio.com/DefaultCollection/WindowsAI/_git/onnxruntime into user/jeffbloo/MergeGithubMasterToDmlDev 2020-04-13 08:39:09 -07:00
Jeff
e3c9fb9fee Fix build issues 2020-04-13 08:36:14 -07:00
Jeff Bloomfield
4e045d39d3 Merge branch 'master' into user/jeffbloo/MergeGithubMasterToDmlDev 2020-04-13 06:36:35 -07:00
Tianlei Wu
54bbbb78ae
Change mask_index input of Attention op to be optional (#3459) 
Change Mask Index to optional
 2020-04-12 22:55:37 -07:00
Sergii Dymchenko
bf3df41424 Put back SubmoduleCheckoutMode parameter into mac-ci.yml. 2020-04-12 21:49:38 -07:00
Sergii Dymchenko
b670cdc11d Remove usage of DeviceProp (which is removed in ort_training) from cudnn_common.cc. 2020-04-12 21:40:54 -07:00
Sergii Dymchenko
7b2fc196fe Get cudnn_common.cc from master. 2020-04-12 21:10:36 -07:00
Sergii Dymchenko
571a6d52f5 Add to list of failing backend tests from master. 2020-04-12 19:16:53 -07:00
George Wu
7f6e407e09
fix python packaging manylinux1 build break. (#3482) 2020-04-11 06:58:22 +08:00
edgchen1
cffdff6702
Publish unit test results from Linux and Mac builds (#3480) 
* Added publish test results step to Linux and Mac builds.

* Fix test result file pattern.
 2020-04-10 14:51:56 -07:00
Tixxx
f5ba9c922d
fix internal loss scale (#3483) 
* Changed internal loss scale to 1-D

* added test

Co-authored-by: root <root@525204a066204ea794f942530b05ae7f000000.axlncovkyjne5caro2tmz3zryb.xx.internal.cloudapp.net>
 2020-04-10 14:13:48 -07:00
Sergii Dymchenko
4b5f66ad17 Put dropout_default, dropout_random, celu back in the list of broken tests. 2020-04-10 13:27:00 -07:00
Ryan Lai
4223591043
Add automatic generation of tensors for Onnxruntime Perf Runner (#3448) 
* Add flag to enable automatic generation of input for models with tensor inputs

* change wording of variable

* Naming convention changes to variables

* Handle free dimensions

* Comment with default allocator

* variable rename

* Remove input_count

* Cast to size_t to avoid warning

Co-authored-by: Ryan Lai <ryalai96@gamil.com>
 2020-04-10 11:54:17 -07:00
stevenlix
56e85484ba
Handle optional inputs and remove more empty shape nodes in TensorRT EP (#3455) 
* check optional inputs and remove more empty shape affected nodes

* fix some minor issues

* update code according to feedback
 2020-04-10 11:13:38 -07:00
edgchen1
20c7dd9f5c
Remove orttraining/docker directory. (#3476) 
The docker images are not publicly available yet.
Addressing PR comment: https://github.com/microsoft/onnxruntime/pull/3174#discussion_r390761308
 2020-04-10 09:41:22 -07:00
Vincent Wang
03996c7c08
Fixes for Where, ConcatGrad and ReduceSumGrad (#3415) 
* Fixes for Expand, Where, ConcatGrad ReduceSumGrad.

* Roll back expand, fix, add tests for reduce grad.

* Roll back CPU Expand change.

* Fix after merge.

Co-authored-by: Vincent Wang <weicwang@microsoft.com>
 2020-04-10 19:35:32 +08:00
Tiago Koji Castro Shibata
d09d4a6b0d
Fix OS build (#3481) 2020-04-09 21:46:01 -07:00
Pranav Prakash
95ade8f47b
Add check to prevent storing nullptr in value_info_ when proto has unused value info (#3461) 
* Add unit test for serialization of unused value_info

* Do not add non-existent (nullptr) value_info_ when loading a model.

Fixes #3430
 2020-04-09 19:25:10 -07:00
Pranav Sharma
2ccedb7b4d
Improve error logging when a kernel cannot be found. (#3473) 
* Improve error logging when a kernel cannot be found.

* Fix mac build
 2020-04-09 19:24:46 -07:00
Sergii Dymchenko
bb2f427990 Remove FastGelu from activations. 2020-04-09 19:17:54 -07:00
KeDengMS
739c9d4875
Always call cudaSetDevice at the beginning of session::Run (#3475) 
This is required for running multithreaded with multi-GPUs. Without it, when running in a work thread it would default to GPU 0, while CUDAExecutionProvider is assigned on other GPUs. That might cause CUDA crash when some CUDA resources is from GPU 0, while being used in GPU N>0.
 2020-04-09 18:54:58 -07:00
Sergii Dymchenko
507d2bb9b9 Get onnxruntime/core/providers/cuda/math/matmul_integer.cc from ort_training. 2020-04-09 18:09:46 -07:00
Sergii Dymchenko
1b465ba820 Get onnxruntime/core/providers/cuda/cu from ort_training. 2020-04-09 18:03:17 -07:00