saymrwulf/onnxruntime
1
0
Fork 0
mirror of https://github.com/saymrwulf/onnxruntime.git synced 2026-07-07 04:39:07 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
7863 commits 2856 branches 57 tags 1.3 GiB
Commit graph

7863 commits

This branch
This branch
All branches
Author SHA1 Message Date
George Wu
355f39ddee
fix cuda build for cmake >= 3.17.0 (#3362) 2020-03-30 00:38:57 -07:00
ytaous
d8f0a0f223
Address PR comments (#3352) 
* PR comments

* revert code for a couple comments

* add negative test case

Co-authored-by: Ethan Tao <ettao@microsoft.com>
 2020-03-29 12:34:54 -07:00
Weixing Zhang
1bbc421884
Don't cast to fp16 in LayernormGrad (#3328) 
Co-authored-by: Weixing Zhang <wezhan@microsoft.com>
 2020-03-28 19:07:32 -07:00
Yang Chen
33b5010e62
skip optional inputs for scan subgraphs (#3349) 
* skip optional inputs for scan subgraphs

We may have cases where the subgraph has optionial inputs that appear
in both subgraph's input and initializer, but not in the node's input.
In such cases, the input model might be invalid, but let's not choke
on it. Instead, let's issue a warning, skip the optional inputs,
and keep going forward.

* address CR feedback
 2020-03-28 16:15:45 -07:00
Dwayne Robinson
324166b9cd Merged PR 4471419: DML EP kernels in ORT 
- Add updated registrations for v11 of existing kernels (ScatterElements, Clip...).
- Add new kernels for GatherElements, GatherND, ScatterND, IsInf, Round, BitShift, CumSum, ReverseSequence, Mod.

Related work items: #23106898, #24655666
 2020-03-28 03:00:19 +00:00
Dwayne Robinson
5bc81b7ae9 Fix bad merge (caused Slice to fail). 2020-03-27 19:58:41 -07:00
Dwayne Robinson
5972ce5566 PR feedback 2020-03-27 19:32:47 -07:00
Dwayne Robinson
f1a062c292 PR feedback 2020-03-27 19:22:25 -07:00
Sherlock
ffb2a3359e
Implement WhereGrad (#3343) 2020-03-27 19:10:40 -07:00
Dwayne Robinson
351c3c30fb Merge branch 'DmlDev' into user/dwayner/DmlEpGatherScatterReverseRangeInfModRoundBitshiftCumSumClip 2020-03-27 18:54:30 -07:00
Dwayne Robinson
8ff351ecc4 Merged PR 4482575: DML EP Slice in ORT 
Related work items: #24672220
 2020-03-28 01:24:44 +00:00
Dwayne Robinson
6c960a9417 PR feedback. 2020-03-27 18:10:46 -07:00
Tiago Koji Castro Shibata
c3cea486d0
Port ConcurrencyTests from TAEF (#3086) 
* Add ConcurrencyTests

* Make ConcurrencyTests compatible with TAEF

* Use test PCH in concurrency tests

* Fix include header

* Ignore unused code warnings on WINML_SKIP_TEST

* Remove BOM

* Remove conflicting namespace in older SDK

* Refactor duplicate code

* Fix unused DELAYLOAD

* Fix unused DELAYLOAD

* Remove link to internal bug

* Address code style fixes

* Add new concurrency tests
 2020-03-27 17:39:22 -07:00
Tixxx
49e6043d07
support Huggingface's adamw (#3318) 
* add weight decay mode to support both pytorch and huggingface's adamw
 2020-03-27 08:04:27 -07:00
Dwayne Robinson
031647635b Delete another litter file. 2020-03-27 02:43:11 -07:00
Dwayne Robinson
5feb3c0f19 Delete litter backup files. 2020-03-27 02:42:09 -07:00
Dwayne Robinson
89df6abac7 Fix slice. 2020-03-27 02:25:58 -07:00
Dwayne Robinson
ccb840ac99 Fix slice. 2020-03-27 00:48:06 -07:00
Yang Chen
5278f73202
Fixed two issues in symbolic_shape_infer script (#3332) 
* Fixed two issues in symbolic_shape_infer script

This change addressed #3293

There were two issues in the script:

* We need to handle a special case for infer_Reshape, where input_shape
is empty and target shape_value is [-1]. In such case, we need to
get sympy data for the output dim (or create one if it doesn't exist).

* We need to update computed dims for newly-created shape for Range op

* also call _update_computed_dims for _infer_Expand

addressed CR feedback

* added ai.onnx into opset list

* instead of manipulating _infer_Reshape, call _update_computed_dims
from _infer_Expand to update newly-computed dims
 2020-03-26 23:27:37 -07:00
Xiang Zhang
810a10b230
Enable Onnxruntime Telemetry by Default for 1.3 (#3338) 2020-03-26 20:57:39 -07:00
ytaous
131c65d23d
Addressing PR comments (#3334) 
* PR comments

* PR comments

* PR comments

* error out bad shape

Co-authored-by: Ethan Tao <ettao@microsoft.com>
 2020-03-26 18:43:30 -07:00
Faith Xu
2e875f4e67
Delete outdated page (#3320) 2020-03-26 18:24:02 -07:00
Pranav Sharma
497e83eda5
Minor update to the issue template. Add a line to attach model where applicable. (#3339) 2020-03-26 14:28:27 -07:00
Xueyun Zhu
0a6ec0df56
Merge pull request #3285 from microsoft/xuzhu/merge_from_master 
Merge from master to ort_training
 2020-03-26 12:10:13 -07:00
Sherlock
d143b41b81
Expose frozen_weights in PyTorch Frontend (#3317) 2020-03-26 11:26:54 -07:00
Hector Li
0e81962e98
correct the cmake version to 3.13 for Arm build (#3333) 2020-03-26 10:20:18 -07:00
Dwayne Robinson
13dabd97b6 Slice 2020-03-25 21:48:14 -07:00
ytaous
66c7579c93
address PR comments (#3312) 
* address PR comments

* PR comments

* PR comments

* disable logging

* typo

Co-authored-by: Ethan Tao <ettao@microsoft.com>
 2020-03-25 19:35:12 -07:00
Changming Sun
5f6ec8ea6d Fix a bug in Maxpool v8 2020-03-25 16:27:43 -07:00
Scott McKay
dee4fc8b8a
Apply the same check for no_transpose from the Reduce* ops to ArgMin and ArgMax (#3315) 2020-03-26 07:41:16 +10:00
Sheil Kumar
51e95ea946
Make ort errors appear in winml exceptions (#3316) 
Co-authored-by: Sheil Kumar <sheilk@microsoft.com>
 2020-03-25 12:20:40 -07:00
Xueyun Zhu
a031af8cfc fix windows build 2020-03-25 19:18:50 +00:00
Xueyun Zhu
9eb792a5b3 move env to .cc file 2020-03-25 16:57:05 +00:00
Wei-Sheng Chin
b38fc0d541
Add bias correction in Adam & Lamb for C++ frontend & python frontend (#3301) 2020-03-25 09:46:44 -07:00
Dwayne Robinson
0f6ace2c36 Fix comments with stale todo's. 2020-03-25 00:16:08 -07:00
Dwayne Robinson
55d32085f6 Update comments. 2020-03-24 23:37:37 -07:00
Dwayne Robinson
cc095fefbb Fix squeeze. 2020-03-24 23:24:58 -07:00
Xueyun Zhu
e9877850a4 fix python error 2020-03-25 01:59:37 +00:00
Scott McKay
4db01309cb
Use GEMM for SVMRegressor. (#3305) 2020-03-25 11:49:44 +10:00
Dwayne Robinson
5366273110 Fix GatherND and ScatterND. 2020-03-24 18:02:39 -07:00
Bowen Bao
6474801ceb
Update ort_trainer.py with lazy onnx export (#3244) 
* Delay onnx export to avoid extra info

* handle cases where onnx model is provided at initialization

* address comments

* fix rebase error
 2020-03-24 13:34:15 -07:00
Tianlei Wu
19edad132c
Move AzureML Bert notebook from onnx tutorial (#3302) 2020-03-24 12:31:02 -07:00
Weixing Zhang
fef7989866
Replacing CudaAsyncBuffer with TArray to improve perf (#3303) 
* removing using CudaAsyncBuffer

* Keep CudaAsyncBuffer for these ops: non_max_suppression, cudnn_rnn_base, concat, split

* fix windows build error

* fix windows build error.

* fix build error

* fix windows build error

Co-authored-by: Weixing Zhang <wezhan@microsoft.com>
 2020-03-24 12:13:27 -07:00
Li-Wen Chang
98c28060b0
Aggregated Send/Recv (#3232) 
* Aggregated Send/Recv

* fix typos

* CR refine

* CR refine

* CR refine

* Add scalar check.

* typo

* reformat

* CR refine

* Forgot to swap order in the implementation after spec changed

* CR refine

* Cr refine

* add Send's input type checking
 2020-03-24 10:20:11 -07:00
Dwayne Robinson
271126a86e Fix clip. 2020-03-24 01:04:08 -07:00
Dwayne Robinson
1df264f4b3 Fix Range. 2020-03-24 00:03:08 -07:00
Hariharan Seshadri
ef7b98f988
Support DisposableNamedOnnxValue inputs in c# Run() (#3175) 
* Initial commit

* Update error message

* Update

* Updates to support holding onto onnxValue and pinnedmemoryBuffer

* Updates

* Minor updates

* Comment out a portion of the tests

* PR feedback

* Minor nit update

* Resolve comments

* PR feedback

* PR updates

* PR feedback
 2020-03-23 18:36:12 -07:00
Faith Xu
fb5ab858d2
Update BUILD instructions (#3282) 
Include guidance for building release packages per question from #3251
 2020-03-23 18:35:22 -07:00
KeDengMS
d15c74e713
Implement pipeline event generator (#3206) 
Implement pipeline event generator with OneFWOneBW schedule in timeline. Each stage of pipeline contains FW and BW of a subset of the model and are scheduled in one worker thread for each microbatch.
 2020-03-23 17:32:54 -07:00
Xueyun Zhu
8f7bd51f7a fix pybind issue introduced by merge 2020-03-23 23:23:34 +00:00