Skip to content

N
ngraph
Switch branch/tag
  1. 07 Feb, 2020 2 commits
    • Robert Kimball's avatar
      Allow topological sort in Function to be replaced (#4206) · 2da03f8a
      Robert Kimball authored 
      * Add replaceable topological sort to Function

* Cleanup

* Cleanup unit test

* Address review comment

* Fix missed item in merge
Co-authored-by: 's avatarScott Cyphers <diyessi@users.noreply.github.com>
      2da03f8a
    • Adam Osewski's avatar
      Use GroupConvolutionBackpropData v1 in ONNX ConvTranspose. (#4158) · d80c8e42
      Adam Osewski authored 
      * Use ConvolutionBackpropData in ONNX ConvTranspose operator.

* Replace deprecated functions and update documentation.

* Remove GroupConvolutionTranspose fused op.

* Add more UT for GroupConvolutionBackpropData.

- They're transformed from GroupConvolutionTranspose

* Fix validate_and_infed types for ConvolutionBackpropData v1.

* Add new constructor for GroupConvolutionBackpropData

- Additionally do some code formatting.

* Improve validation in GroupConvBackprop v0.

* Update/fix upgrade & downgrade passes for
(Group)ConvolutionBackpropData.

* Add decomposition and validation for GroupConvolutionBackpropData:v1

* Type properties UT for GroupConvolutionBackprop v1

* Remove include of deleted header.

* Add squeeze to opset1::builder.

* Use opset1 squeeze.

* Helper functions for padding and output spatial shape inference.

These helper functions are specialized only for *ConvolutionBackpropData operators.

* Update documentation.

* Add more checks and remove deprecated assert.

* Refactor GroupConvolutionBackpropData validation.

- Use helper functions to infer padding and output shape.
- Change get_output_shape to return only spatial features dimensions

* Remove unnecessary code from donwgrade pass.

The GroupConvolutionBackpropData node has already computed all those required
information.

* Call Node API. Small review suggestions.

* Work only on spatial dimensions inside helper functions.

* Enhance documentation for ConvolutionBackpropData

* Refactoring.

* Reuse values computed in downgraded operator.

* Check for correct filters input shape.

* Add additional check on auto pad value.

* Remove unused variables.

* Rename helper function.

* Add more checks.

* Override is dynamic and add more restrictive check.

* Fix merge artifact and force unsigned type for argument.

* Review comments: renaming.

* Change Convolution + bias pattern.

Change from bias broadcasting to appropriate convolution output shape
into reshape into (1, C, 1, 1) shape, which sould work better.

* Add few more test cases for ONNX ConvTranspose op.

* Skip one UT failing on PlaidML backend due to accuracy.
      d80c8e42
  3. 06 Feb, 2020 1 commit
  5. 04 Feb, 2020 2 commits
  7. 03 Feb, 2020 3 commits
  9. 02 Feb, 2020 1 commit
  11. 31 Jan, 2020 3 commits
    • Scott Cyphers's avatar
      Printing cleanup (#4223) · 02a8fa95
      Scott Cyphers authored 
      * Printing cleanup

Add Input/Output
Simplify element type
Simplfy Node printing  implementation, try to de-noise it a bit
Enable printing of Node*

* Adjust printing

* Add doc note

* Update src/ngraph/node.cpp
Co-Authored-By: 's avatarRobert Kimball <robert.kimball@intel.com>

* Update src/ngraph/node.hpp
Co-Authored-By: 's avatarRobert Kimball <robert.kimball@intel.com>

* Cleanup

* typo
Co-authored-by: 's avatarRobert Kimball <robert.kimball@intel.com>
      02a8fa95
    • Mateusz Bencer's avatar
      Extend NgraphTestCase to support dynamic backends (#4234) · e37b386e
      Mateusz Bencer authored 
      * Switch to PartialShape in onnx_importer ValueInfo

* Construct dynamic dimensions out of ONNX dimensions defined as dim_param

* Validate the PartialShape of inputs created from an ONNX model with dynamic shapes

* Validate the output shape inference for a dynamic ONNX model

* Test the execution of an ONNX model with dynamic dimensions

* Test the Ax+B with more than one batch size

* Provenance tagging adjustments - PartialShape instead of Shape

* Correct translation of ONNX shapes to nG shapes

* Test the shape of Constant produced by scalar initializers

* Review comments & more strict assertions in UT

* UT checking a dynamic rank input

* Fully dynamic input inference test

* first dynamic version

* modified UTs

* Added assert checks

* Added specialised methods

* first verion of AvgPool

* code review remarks introduced

* Changed tests to use default BackendMode value

* Reverted not related changes

* code review remarks introduced

* Set static backend as default
Co-authored-by: 's avatarTomasz Dołbniak <tomasz.dolbniak@intel.com>
Co-authored-by: 's avatarSang Ik Lee <sang.ik.lee@intel.com>
      e37b386e
    • Robert Kimball's avatar
      Allow Constant ops to share internal buffer (#4216) · 2c23cf20
      Robert Kimball authored 
      * Move non-templated constructor implementation to the source file

* Optimize constant constructor for uniform constant

* Cleanup

* Much faster deserialize constant

* Adding unit tests

* Unit tests

* Update unit test

* Cleanup

* style

* Cleanup nbench output

* wip

* Fix specializations

* Change from unique to shared_ptr internally

* Enable copy of Constant

* cleanup Constant ctors

* Fix copy contructor

* Fix compile error
Co-authored-by: 's avatarScott Cyphers <diyessi@users.noreply.github.com>
      2c23cf20
  13. 30 Jan, 2020 2 commits
  15. 29 Jan, 2020 2 commits
  17. 28 Jan, 2020 1 commit
    • Nishant Patel's avatar
      LRU cache for dynamic shapes (#3827) · 44b58722
      Nishant Patel authored 
      * LRU cache for caching graphs

* static

* LRU Cache

* Cache graph

* Make data members static

* Temp test case functional

* Temp test case functional

* Fix shape

* Make lru local to dynamic wrapper

* Make cache thread safe

* Remove static copies of data members

* Add a separator  between shapes of different inputs

* Clear list and map in destructor

* Caching on values of shape relevant inputs

* Replace cout's by NGRAPH_INFO

* Add a environment variable for cache size

* Add mutex header

* style

* change to int64_t

* Save the cloned function to get the output shape to allocate output storage

* Pass inputs without wrapping

* Fix conv shape relevant inputs

* gcc 4.8 doesnt support ostring stream as a copyable object

* Pass key by reference

* PR feedback

* Apply suggestions from code review

* Replace malloc
Co-authored-by: 's avatarScott Cyphers <diyessi@users.noreply.github.com>
Co-authored-by: 's avatarRobert Kimball <robert.kimball@intel.com>
      44b58722
  19. 27 Jan, 2020 1 commit
  21. 23 Jan, 2020 4 commits
  23. 22 Jan, 2020 1 commit
  25. 21 Jan, 2020 1 commit
  27. 17 Jan, 2020 2 commits
    • Sang Ik Lee's avatar
      Ninja build (#4101) · 6e415019
      Sang Ik Lee authored 
      * Add new option to do a clean Ninja build.

* Enable Interpreter and unit-test.

* Enable json.

* Enable NOP backend.

* Enable tools.

* Enable CPU DEX ONLY on Linux.

* Quick hack to enable dex only on Mac.

* Enable onnx import.

* Cleanup gtest.

* Cleanup tbb_prebuilt.

* Fix typo.

* Move gtest include point.

* Fix gtest windows path.

* Remove temporary option.

* Keep GCPU upto date with INTERPRETER.

* Add missing file.

* Enable GCPU.

* Enable CPU codegen.

* Remove tabs.

* Revert "Merge branch 'silee2/gcpu' into silee2/ninja"

This reverts commit e513cd3fe3659a477af6e8fab6c1852a5b1493c3, reversing
changes made to 31dca3301cd89e15af23acd0d50aa06a1c5d6001.

* Enable PlaidML.
Co-authored-by: 's avataraslepko <44713115+aslepko@users.noreply.github.com>
      6e415019
    • Nagy Mostafa's avatar
      [MLIR] Enable lowering of GroupConv in MLIR CPU backend (#4102) · 8e46ff86
      Nagy Mostafa authored 
      * WIP

* WIP

* Refactored existing convolution

* Add Channel and num of filters bounds parameters to helper

* Works on unit-tests. v1 op gets converted and breaks

* Fixed group conv with groups in filters shape. Tests pass

* style

* add LIT tests

* Switch outer loop to affine loop

* re-org code

* PR fixes

* Revert ops.td

* PR fixes
      8e46ff86
  29. 16 Jan, 2020 3 commits
  31. 14 Jan, 2020 3 commits
    • Robert Kimball's avatar
      Revert "Remove conditional compile (#4106)" (#4178) · c748bda4
      Robert Kimball authored 
      This reverts commit 1e58565a.
      c748bda4
    • Tomasz Dołbniak's avatar
      [ONNX] Extended support for provenance tags (#4154) · 9bfbd3c6
      Tomasz Dołbniak authored 
      * Checking if provenance_tags key exists

* Add provenance tag prototype

* Format provenance tag

* Display provenance tag

* Clean debug printing

* Add const to variables

* Separate method for add provenance tags

* Return NodeVector reference

* Return const NodeVector

* Moved add_provenance_tags function to commons

* Style apply

* Simple model for tests

* Provenance tag test

* Expect substring instead of  equal

* Add provenance tags to intermediate nodes recursively

* One tag per node

* Add traverse node args instead of recursion

* Return NodeVector instead of set of pointers

* Use treverse_nodes and lambda function

* Remove unused helper functions

* Remove is_constant() condition

* Update test model prototxt

* Update test substring

* Use node name and output names to build provenance tags in onnx importer

* Unit tests for onnx_importer provenance tags

* Missing <numeric> include

* Add provenance tags to constants buit from ONNX initializers

* Add provenance tags to Constants and Parameters created out of ONNX inputs and initializers

* More strict assertions in onnx provenance tests

* Unit test for onnx importer Parameter nodes tagging

* Helper function for the onnx provenance tests

* Some docs

* Obsolete comment removal

* Separate file for onnx provenance tags unit tests

* Code formatting

* Move the inputs tagging to the Graph class

* Tagging moved to the Graph class entirely

* Missing include and extra helper variable

* Unit tests helper documentation

* Change the UT helper to lowercase
Co-authored-by: 's avatarKatarzyna Mitrus <katarzyna.mitrus@intel.com>
      9bfbd3c6
    • Nagy Mostafa's avatar
      [MLIR] In-place memory optimization for elt-wise and concat ops. (#3832) · 6de4893b
      Nagy Mostafa authored 
      * AliasMap WIP

* Added liveness info

* WIP

* WIP: Tests

* WIP: LIT tests

* Added knobs for mem optimization pass. More LIT tests

* Revert affine_lowerer change

* More elaborate comment

* Minor fixes

* style-apply

* Rename liveness

* Switch to Analysis framework

* Fix optimization conditions

* Remove LIT test

* style

* Switch to equivalence relationship impl of non-alias relationship

* refined comment

* Switch non-alias to equivalence relationship

* Fix bad merge

* Adding tests. WIP

* Added buffer size tracking and unit-tests

* Added LIT and unit-tests

* Turn optimization ON

* style

* fix unit-tests

* Fix useCount

* Fix copyright and typo

* Refine few comments, remove new lines

* style fix
Co-authored-by: 's avatarScott Cyphers <diyessi@users.noreply.github.com>
Co-authored-by: 's avatarSang Ik Lee <sang.ik.lee@intel.com>
      6de4893b
  33. 13 Jan, 2020 1 commit
  35. 11 Jan, 2020 3 commits
  37. 10 Jan, 2020 4 commits