Narumol Prangnawarat [Fri, 30 Oct 2020 16:06:55 +0000 (16:06 +0000)]
IVGCVSW-5322 Fix segfault between Neon and Cl layers
* Fallback to memory copy if memory import is not supported
* Remove direct compatibility between Neon and Cl Tensors
* Unit tests fallback from Neon to Cl and Cl to Neon
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
Change-Id: Iec00a77423fb23b37a6b1aefee1b2ec4d649efca
Colm Donelan [Fri, 30 Oct 2020 14:46:21 +0000 (14:46 +0000)]
IVGCVSW-5265 Removing more Boost references from test executables.
* Removed unused includes from InferenceModel.hpp.
* Replaced use of boost multi-array with vectors in YoloInferenceTest.
Signed-off-by: Colm Donelan <Colm.Donelan@arm.com>
Change-Id: Ieadf3471ed170b09859187c83616c8e249f94543
James Ward [Fri, 30 Oct 2020 12:36:19 +0000 (12:36 +0000)]
IVGCVSW-5267 Remove boost from core android-nn-driver
Signed-off-by: James Ward <james.ward@arm.com>
Change-Id: I19088511be05087f979aa285242dd18f093fc5da
James Ward [Fri, 30 Oct 2020 09:54:21 +0000 (09:54 +0000)]
IVGCVSW-5266 Remove Boost from standalone dynamic backend
Signed-off-by: James Ward <james.ward@arm.com>
Change-Id: I11e02826dd155bc722d6659d9b7e3053cad45b7f
James Ward [Thu, 29 Oct 2020 16:19:02 +0000 (16:19 +0000)]
IVGCVSW-5468 Rewrite QuantizationDataSet.cpp to avoid use of CsvReader
* Remove armnnUtils/CsvReader and usage
* Remove armnn/CsvReaderTest and usage
* Replace functionality in QuantizationDataSet.cpp
Signed-off-by: James Ward <james.ward@arm.com>
Change-Id: I7213904482afa93ae6d607aa5e69117c8c34ea81
Sadik Armagan [Thu, 29 Oct 2020 16:14:54 +0000 (16:14 +0000)]
IVGCVSW-5379 'TfLiteDelegate: Implement the ElementWiseBinary operators'
* Implemented ADD operator
* Implemented FP32 unit tests for ADD operator
Signed-off-by: Sadik Armagan <sadik.armagan@arm.com>
Change-Id: Id7238749308855bd2b2118f4b6e60e765815c38f
James Ward [Wed, 21 Oct 2020 08:32:46 +0000 (09:32 +0100)]
IVGCVSW-5267 Remove boost from core android-nn-driver
* WIP
!armnn:4246
Signed-off-by: James Ward <james.ward@arm.com>
Change-Id: I1b379b9e7f4397ca9b9189e423042469af382c09
Colm Donelan [Thu, 29 Oct 2020 11:39:14 +0000 (11:39 +0000)]
IVGCVSW-5265 Remove boost from core ArmNN CMake
* Remove all but Boost_UNIT_TEST_FRAMEWORK_LIBRARY from
ArmNN Cmake files.
* Remove references to boost::fpc from old test applications.
Signed-off-by: Colm Donelan <Colm.Donelan@arm.com>
Change-Id: Ibb1261dee4b971d1788d2805528aa800a8b883ce
Teresa Charlin [Thu, 15 Oct 2020 12:16:07 +0000 (13:16 +0100)]
IVGCVSW-5314 Create OptimizeForExclusiveConnection
* FuseBatchNorm class has been added to facilitate testing
* Only Convolution2D FP32 being fused
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: I049c4770946ddca21b08516d4c9f4d0d22bf9b45
Nikhil Raj [Thu, 29 Oct 2020 12:25:10 +0000 (12:25 +0000)]
Update ACL pin to
5a4284dc7d98a382d0fa492b64fabe430d5afdc6
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: Ife58d18880bc561bede5cd432befb632fe9dc9c9
Nikhil Raj [Wed, 28 Oct 2020 12:54:19 +0000 (12:54 +0000)]
IVGCVSW-4995 Update BuildGuideAndroidNDK and BuildGuideCrossCompilation md files in ArmNN
* Updating the docs with latest changes made to protobuf, caffe and onnx
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: I336ccf13ab38629399fdbdf70479d5fb07fa43fa
Teresa Charlin [Tue, 27 Oct 2020 16:53:09 +0000 (16:53 +0000)]
Update ACL to
19a41bad7fbbc18fc9032753b234b1f7c632b2d5
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: Ia2388b54d72764ead8818e559843f46f8d76e245
Pavel Macenauer [Thu, 22 Oct 2020 08:07:59 +0000 (08:07 +0000)]
Move PyArmNN unit test resources for IDeserializer to external storage
Change-Id: I582acd5a852143b17ca19aebaf54ccf5c7b5df87
Signed-off-by: Pavel Macenauer <pavel.macenauer@nxp.com>
Finn Williams [Thu, 22 Oct 2020 15:53:35 +0000 (16:53 +0100)]
IVGCVSW-5433 Remove boost::transform_iterator and make_transform_iterator
Signed-off-by: Finn Williams <Finn.Williams@arm.com>
Change-Id: I28aace7092cff5743353df1b1de8e7a4691554d3
Francis Murtagh [Tue, 27 Oct 2020 16:53:25 +0000 (16:53 +0000)]
Bugfix: Don't try to link pipeCommon if it's not being built
* If BUILD_TIMELINE_DECODER=0 it gave /usr/bin/ld: cannot find -lpipeCommon
Signed-off-by: Francis Murtagh <francis.murtagh@arm.com>
Change-Id: Ib10e894c54f32370a96b702a6f1b462fda31388b
Narumol Prangnawarat [Tue, 27 Oct 2020 18:07:04 +0000 (18:07 +0000)]
Fix BackendHint missing when cloning a layer
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
Change-Id: I550fb59469af39f8fa3415843e973f06b18485e5
Sadik Armagan [Tue, 27 Oct 2020 17:30:18 +0000 (17:30 +0000)]
IVGCVSW-5378 'TfLiteDelegate: Implement the ElementWiseUnary operators '
* Moved ElementwiseUnary operators tests into single file
* Implemented FP32 test for supported ElementwiseUnary operators
Signed-off-by: Sadik Armagan <sadik.armagan@arm.com>
Change-Id: I4b7eab190c3c8edb50927b8e1e94dd353597efcb
Francis Murtagh [Tue, 27 Oct 2020 15:20:40 +0000 (15:20 +0000)]
Bugfix: Allow use of dynamic backendId in execute network
* Stops execute network rejecting backendId based on BackendRegistry
* Dynamically loaded backends arent visible yet as runtime isn't initialized
Change-Id: I87adfd137b2225ab07f8c3e996db9565caf276eb
Signed-off-by: Francis Murtagh <francis.murtagh@arm.com>
Sadik Armagan [Fri, 23 Oct 2020 16:14:43 +0000 (17:14 +0100)]
IVGCVSW-5366 'Add a do nothing SubGraph class'
IVGCVSW-5373 'Implement the ABS operator in the Delegate'
* Added a Switch statement into the VisitNode() function
* Separated the Visit functions into the categorized source files
* Implemented VisitElementwiseUnary() function
* Added tests for ABS and SQRT
Signed-off-by: Sadik Armagan <sadik.armagan@arm.com>
Change-Id: If9654d0a8d8ff7dcd6fb5cbe0dc312941772affb
Teresa Charlin [Mon, 26 Oct 2020 18:55:11 +0000 (18:55 +0000)]
Update ACL pin to
9ae06d4986bc3055f7786c1097b465bd321cf8eb
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: Ie9e0e1e3096edbbff7cf05fe65e8eeeb819d59bf
Nikhil Raj [Fri, 9 Oct 2020 13:52:25 +0000 (14:52 +0100)]
IVGCVSW-5077 Ensure ArmNN builds successfully with latest protobuf
* Use the single parameter version of SetTotalBytesLimit()
* Update CMakeLists to turn off deprecated declarartions
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: I2d360966743986872cfef40c2ab1a3505fc5d99a
Nikhil Raj [Wed, 7 Oct 2020 09:31:06 +0000 (10:31 +0100)]
Update ACL pin to
b333758b5e2b00ba7b36cc02c169605f55fb0251
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: I2280e7fd77729c41086f7994fde182255159d83d
antkillerfarm [Thu, 15 Oct 2020 03:02:07 +0000 (11:02 +0800)]
GitHub#465 Fix NonMaxSuppression
If visited flag set true, it should not be visited any more.
For example, if we put 10 boxes (ordered by score) into NonMaxSuppression:
* Step1: Suppose Box 2/3/6/8 are suppressed by Box 1. Box 4/5/7/9/10 survived.
* Step2: Correct way: We use Box 4 to suppress the survive boxes.
Prior to this commit: Box 4 may be suppressed by Box 2,
even Box 2 is already suppressed by Box 1...
Signed-off-by: Antkillerfarm <antkillerfarm@gmail.com>
Change-Id: I38d7a84287649827a16565748592fb562b4df5d5
wangg [Wed, 26 Aug 2020 01:44:32 +0000 (01:44 +0000)]
Add IDeserializer support in pyarmnn
Resources required for new unit tests are included for review.
Signed-off-by: Guanqun Wang gemini910621@gmail.com
Change-Id: Iead6cb5beaf824a6f467ad9da4aede5719ebe4ec
Jan Eilers [Thu, 15 Oct 2020 17:34:43 +0000 (18:34 +0100)]
IVGCVSW-5284 Refactor ExecuteNetwork
* Removed boost program options and replaced it with cxxopts
* Unified adding, parsing and validation of program options
into the struct ProgramOptions
* Program options are now parsed directly into ExecuteNetworkParams
which can be passed directly to MainImpl
* Split NetworkExecutionUtils into header and source
* Removed RunTest
* Removed RunCsvTest
* Removed RunClTuning
* Moved MainImpl back to ExecuteNetwork.cpp
* Added additional util functions
The functionality of ExecuteNetwork remains the same. Only
cl tuning runs need to be started separately
and there is no short option for fp16-turbo-mode because -h is
reserved in cxxopts to print help messages
Signed-off-by: Jan Eilers <jan.eilers@arm.com>
Change-Id: Ib9689375c81e1a184c17bb3ea66c3550430bbe09
Sadik Armagan [Mon, 19 Oct 2020 16:35:30 +0000 (17:35 +0100)]
IVGCVSW-5365 'Create the TfLite Delegate subdirectory in ArmNN'
* Created delegate sub-directory under armnn
* Created Delegate, ArmnnSubgraph and DelegateOptions classes
* Created cmake files.
* Integrated doctest (under MIT license) as testing framework
Signed-off-by: Sadik Armagan <sadik.armagan@arm.com>
Change-Id: If725ebd62c40a97c783cdad22bca48709d44338c
Francis Murtagh [Thu, 15 Oct 2020 10:02:12 +0000 (11:02 +0100)]
Bugfix: fix typo in Pyarmnn README example code
Signed-off-by: Francis Murtagh <francis.murtagh@arm.com>
Change-Id: Ib2c219c976dd37e570681442e15e414cdb61b3f6
Matthew Sloyan [Thu, 15 Oct 2020 12:53:27 +0000 (13:53 +0100)]
IVGCVSW-5435 Add FloatingPointComparison to remove boost::math::fpc uses
* Added FloatingPointComparison.hpp and FloatingPointComparisonTest.cpp,
which compares two floats and returns true if the values are
within a specified or default tolerance of each other.
* Also removed boost::math::fpc from test/TensorHelpers.hpp to validate.
Signed-off-by: Matthew Sloyan <matthew.sloyan@arm.com>
Change-Id: I164c32eccd213c53bb1bc4f9cd4ee4838f1781c9
Colm Donelan [Wed, 14 Oct 2020 16:54:39 +0000 (17:54 +0100)]
IVGCVSW-5406 Update README.md with details of third party tools.
* Add Description, version and provenience to third party tools table.
* Update arm-security@arm.com email address.
Signed-off-by: Colm Donelan <Colm.Donelan@arm.com>
Change-Id: Id48f36ae2319688fafea9fc4757bc75b724b3d02
James Ward [Mon, 12 Oct 2020 13:17:36 +0000 (14:17 +0100)]
IVGCVSW-5280 Switch tests/InferenceTest and derived tests over to cxxopts
* refactor AddCommandLineOptions() functions to allow checking of required options
* add CxxoptsUtils.hpp file for convenience functions
!referencetests:268500
Signed-off-by: James Ward <james.ward@arm.com>
Change-Id: Ica954b210b2981b7cd10995f0d75fcb2a2f7b443
Mike Kelly [Wed, 14 Oct 2020 10:48:21 +0000 (11:48 +0100)]
IVGCVSW-5335 Added Documentation for fast_math
* Added Documentation for fast_math to CLBackendModelContext
* Added Documentation for fast_math to NeonBackendModelContext
Signed-off-by: Mike Kelly <mike.kelly@arm.com>
Change-Id: I43a0568ae6914e074a80130a051e5d9bb849f2ba
Jim Flynn [Tue, 13 Oct 2020 13:40:29 +0000 (14:40 +0100)]
IVGCVSW-5434 Remove boost/preprocessor.hpp
Change-Id: I51462d18ce0be4b88a23453cfdd16510f30dd1e3
Signed-off-by: Jim Flynn <jim.flynn@arm.com>
Matthew Sloyan [Mon, 12 Oct 2020 14:03:01 +0000 (15:03 +0100)]
IVGCVSW-4489 Remove remaining occurrence of boost::format
* Replaced with fmt::format in Descriptors.cpp.
* Removed remaining boost/format headers in ArmNN codebase.
* Removed additional boost header in Network.cpp
Signed-off-by: Matthew Sloyan <matthew.sloyan@arm.com>
Change-Id: Ib98b83bf4ec99ef98ce7a3635ec0dd478c3e43e1
Francis Murtagh [Mon, 12 Oct 2020 15:45:49 +0000 (16:45 +0100)]
Bugfix: Correctly parse the armnn_includes by separating with space
* Stops generate_wrap reading as -I/usr/local/include-I/usr/include
* Allows generation using system headers
Signed-off-by: Francis Murtagh <francis.murtagh@arm.com>
Change-Id: I501c473b0f624f80ebccfade7b009a74bdcb2d0c
Mike Kelly [Thu, 8 Oct 2020 11:19:01 +0000 (12:19 +0100)]
IVGCVSW-5335 Documentation for fast_math
* Changed documentation for fast_math to add warning about possibly reduction in precision.
Signed-off-by: Mike Kelly <mike.kelly@arm.com>
Change-Id: If954471efc6aef702e8490585571815d9e19b3fc
Derek Lamberti [Thu, 17 Sep 2020 12:58:18 +0000 (13:58 +0100)]
Load dynamic backends for YoloV3
* Optional cmd option to dump optimized model to dot
* Optional cmd option to specify dynamic backends path
* input is now optional and must exist if given
* comparison files now optional and must exist if given
Change-Id: I1499c9eb715be3cacdba2c227e1a93dd997f355d
Signed-off-by: Derek Lamberti <derek.lamberti@arm.com>
James Ward [Mon, 5 Oct 2020 16:11:23 +0000 (17:11 +0100)]
IVGCVSW-5287 Switch tests/MultipleNetworksCifar10 over to cxxopts
Signed-off-by: James Ward <james.ward@arm.com>
Change-Id: Ia4ff69f2c8dd538ad7845053dc7a690f434c381c
Matthew Sloyan [Thu, 8 Oct 2020 15:41:47 +0000 (16:41 +0100)]
IVGCVSW-5279 Switch armnnQuantizer over to cxxopts
Signed-off-by: Matthew Sloyan <matthew.sloyan@arm.com>
Change-Id: I79f12ba33b3ca58cdc4be531ffbf72fa20690792
David Monahan [Fri, 9 Oct 2020 11:30:04 +0000 (12:30 +0100)]
IVGCVSW-5291 Fix for Yolov3 producing 0s on Neon
Signed-off-by: David Monahan <david.monahan@arm.com>
Change-Id: I9331d590e71fc479979c11be9dc750d6435f12bc
Sadik Armagan [Fri, 9 Oct 2020 12:39:17 +0000 (13:39 +0100)]
IVGCVSW-5173 'Upgrade from v1.15 of tensorflow to v2.3'
* Fixed the tensorflow build issue.
Signed-off-by: Sadik Armagan <sadik.armagan@arm.com>
Change-Id: I608c55fdf76ff4d436434c0a50b5368923d2ddb4
Matthew Sloyan [Mon, 28 Sep 2020 11:58:14 +0000 (12:58 +0100)]
IVGCVSW-5282 Switch tests/TfLiteMobilenetQuantized-Armnn over to cxxopts
* Required update to cxxopts v3.0 for unrecognised options to work.
Signed-off-by: Matthew Sloyan <matthew.sloyan@arm.com>
Change-Id: If1803731474e42580d663c802a1f3ff240fadffe
Jim Flynn [Thu, 8 Oct 2020 10:42:30 +0000 (11:42 +0100)]
IVGCVSW-5363 Add Unmap layer and Unmap workload
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
Signed-off-by: Jim Flynn <jim.flynn@arm.com>
Change-Id: Ie5ecfa67e4763d0c058905592fe2e2fd7315f85c
Matthew Sloyan [Tue, 6 Oct 2020 09:03:21 +0000 (10:03 +0100)]
IVGCVSW-5286 Switch tests/ModelAccuracyTool-Armnn over to cxxopts
* Fixed two code errors which were causing compiler issues.
* Added support for multiple input and output tensors.
Signed-off-by: Matthew Sloyan <matthew.sloyan@arm.com>
Change-Id: I298e4bf52a0b41349adab60295e0f22da3cc057b
Teresa Charlin [Tue, 6 Oct 2020 11:11:39 +0000 (12:11 +0100)]
Remove Resize from list of layers that need padding in Neon
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: I054f0b71d4e9581c637fa09e40f6b661e58e39f3
Matthew Sloyan [Tue, 6 Oct 2020 09:45:32 +0000 (10:45 +0100)]
IVGCVSW-5405 Remove boost::make_iterator_range and boost::to_upper_copy
* Removed from ModelAccuracyTool-Armnn and ImageCSVFileGenerator
* Fixed formatting in ImageCSVFileGenerator
Signed-off-by: Matthew Sloyan <matthew.sloyan@arm.com>
Change-Id: I09dfca27813582cc48f46d0507680368ed823a9c
Jim Flynn [Tue, 6 Oct 2020 09:14:50 +0000 (10:14 +0100)]
IVGCVSW-5362 Add Map layer and Map workload
Signed-off-by: Jim Flynn <jim.flynn@arm.com>
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
Change-Id: Id2227c58809b84c7a7af61f7c0d88ad7d45ce558
Matthew Sloyan [Tue, 6 Oct 2020 15:06:07 +0000 (16:06 +0100)]
IVGCVSW-4488 Update cxxopts to version 3.0
* Required to fix issue in TfLiteMobilenetQuantized-Armnn.cpp
Signed-off-by: Matthew Sloyan <matthew.sloyan@arm.com>
Change-Id: I3a465f62e3d656c9626113da6223e4fa26b535a6
Colm Donelan [Tue, 6 Oct 2020 14:15:33 +0000 (15:15 +0100)]
IVGCVSW-5213 Switch armnnConverter over to cxxopts
* Swap out boost::program_options with cxxopts.
* Remove const from argv declaration to conform to cxxopts.
* Clean up CLI handling around multiple inputs and shapes.
Each input must be specified with its own -i param.
Each corresponding -s parameter must be specified too.
Signed-off-by: Colm Donelan <Colm.Donelan@arm.com>
Change-Id: I01c2683dbca51b2322c98913731ac82a59e400a5
Sadik Armagan [Thu, 1 Oct 2020 16:12:17 +0000 (17:12 +0100)]
IVGCVSW-5173 'Upgrade from v1.15 of tensorflow to v2.3'
* Updated the generate_tensorflow_protobuf script to search for .proto
files in tensorflow directory instead of getting from a file
* Updated BuildGuideAndroidNDK.md to clone Tensorflow 2.3.1
Signed-off-by: Sadik Armagan <sadik.armagan@arm.com>
Change-Id: I8d2b0f0a04327f930784648e9e8f7f25895d53d1
Mike Kelly [Mon, 5 Oct 2020 11:12:49 +0000 (12:12 +0100)]
IVGCVSW-5407 Failing Unittest
* The number of error and warning messages will vary depending on the device but will always be greater than 0.
Signed-off-by: Mike Kelly <mike.kelly@arm.com>
Change-Id: Iee1d12e60d8c88c9116083d7274cea2ec23cb032
Teresa Charlin [Mon, 5 Oct 2020 08:53:18 +0000 (09:53 +0100)]
Update ACL pin to
fc2f6d0427e1d886fcccc68867d1af1ccd96608b
* Set use_padding to false in neon workload
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: Ia5367de331efe1d28dea4dfbefc0713720da81f9
Éanna Ó Catháin [Thu, 10 Sep 2020 12:02:37 +0000 (13:02 +0100)]
PyArmNN Updates
* Updated setup.py to raise error on mandatory ext
* Updated examples section of main readme
* Added readme to img class
* Moved img class to new subdir
Change-Id: Iea5f6d87c97e571b8ca5636268231506538840c7
Signed-off-by: Éanna Ó Catháin <eanna.ocathain@arm.com>
Signed-off-by: Jakub Sujak <jakub.sujak@arm.com>
Colm Donelan [Wed, 9 Sep 2020 11:48:16 +0000 (12:48 +0100)]
IVGCVSW-5297 Remove boost::format from rest of ArmNN.
* Replacing calls to boost:format with fmt:format.
* TensorUtils.cpp added outputShape.reserve call.
Signed-off-by: Colm Donelan <Colm.Donelan@arm.com>
Change-Id: I4b2ed0f72039df824a2adca9309b8a9bbb158c5b
Matthew Sloyan [Wed, 23 Sep 2020 15:57:23 +0000 (16:57 +0100)]
IVGCVSW-5334 Remove remaining boost::numeric_cast from armnn
* Floating point casts now use armnn::numeric_cast.
* Also removed remaining header imports.
Signed-off-by: Matthew Sloyan <matthew.sloyan@arm.com>
Change-Id: I2d37847d67f164fc0a0ae17f34d49ff3d2210c30
Nikhil Raj [Fri, 2 Oct 2020 10:11:43 +0000 (11:11 +0100)]
Update ACL pin to
b84f9d34dbb857ad6113c0c89ad109498fa75fe5
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: I0b8f698f20889eb09b7c881d7daadbd05f8db3c7
Sadik Armagan [Tue, 29 Sep 2020 14:12:36 +0000 (15:12 +0100)]
IVGCVSW-4997 'Superfluous memcopy workloads'
* If Output Layer is already connected to MemCopy Layer do not insert
CopyMemGenericWorkload.
Signed-off-by: Sadik Armagan <sadik.armagan@arm.com>
Change-Id: I9f813be5a3de2bc62d16864edb3eeaf371ef48e0
Matthew Sloyan [Wed, 23 Sep 2020 08:51:04 +0000 (09:51 +0100)]
IVGCVSW-5306 Add floating point type checks to NumericCast.hpp
* Added Unit Tests to capture all combinations.
Signed-off-by: Matthew Sloyan <matthew.sloyan@arm.com>
Change-Id: I04db920a5f5f485dc00b2f16582cf7e0bbef3ef2
James Ward [Thu, 10 Sep 2020 10:57:28 +0000 (11:57 +0100)]
IVGCVSW-5294 Remove boost::format armnn backends
* replaced with fmt::format
* one case required std:stringstream instead
Signed-off-by: James Ward <james.ward@arm.com>
Change-Id: Ife7c4cf5f143e43373f42edf6124158af132abc5
James Ward [Fri, 11 Sep 2020 16:32:44 +0000 (17:32 +0100)]
IVGCVSW-5296 Remove boost::format armnn parsers
* replaced with fmt::format
* one case required std::stringstream instead
Signed-off-by: James Ward <james.ward@arm.com>
Change-Id: Ica9a7eb4e7bed04aa03172058dd9e3d10efc8548
Matthew Sloyan [Fri, 25 Sep 2020 10:43:32 +0000 (11:43 +0100)]
IVGCVSW-5281 Switch tests/ImageCSVFileGenerator over to cxxopts
Signed-off-by: Matthew Sloyan <matthew.sloyan@arm.com>
Change-Id: I1b2ba4fcaebbafb70693b83b40b79db0071b4522
Nikhil Raj [Thu, 1 Oct 2020 09:55:05 +0000 (10:55 +0100)]
Update ACL pin to
de2e747de0a63933d0cfcb75739d9460ea7c2ff5
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: Ie2371f743e2ae87617693052d3887a4e6732accc
Matthew Sloyan [Fri, 25 Sep 2020 16:22:00 +0000 (17:22 +0100)]
IVGCVSW-5285 Switch tests/ImageTensorGenerator over to cxxopts
Signed-off-by: Matthew Sloyan <matthew.sloyan@arm.com>
Change-Id: I9d43e347021634022a875fc4526baa438c467df0
Matthew Sloyan [Wed, 30 Sep 2020 12:07:51 +0000 (13:07 +0100)]
IVGCVSW-5283 Switch tests/profiling/gatordmock over to cxxopts
Signed-off-by: Matthew Sloyan <matthew.sloyan@arm.com>
Change-Id: I4f06a414d6bd5f18c2ced882ac518052ed371cfc
Rob Hughes [Fri, 28 Aug 2020 10:48:35 +0000 (11:48 +0100)]
Include layer GUID in SerializeToDot output
Change-Id: I1a6df60683cc51fcd9739b6dc98f1e722becf045
Signed-off-by: Robert Hughes <robert.hughes@arm.com>
Teresa Charlin [Tue, 29 Sep 2020 07:38:47 +0000 (08:38 +0100)]
COMPMID-3784 Fix 1 CTS MUL INT32 failure due to using SATURATE
* LargeGraph_TENSOR_INT32_Rank4/26
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: I9d07444db56e26c13a77bf022938644ed7953d6b
Finn Williams [Tue, 29 Sep 2020 18:54:00 +0000 (19:54 +0100)]
IVGCVSW-5325 Fix non-channel per axis quantization
Signed-off-by: Finn Williams <Finn.Williams@arm.com>
Change-Id: Ie0cf69b2cd76d6ecedab43d3d9ae267d23bbc052
James Ward [Mon, 28 Sep 2020 10:56:35 +0000 (11:56 +0100)]
IVGCVSW-4519 Remove Boost Variant and apply_visitor variant
* replace boost::variant with mapbox::util::variant
* replace boost::apply_visitor with mapbox::util::apply_visitor
* replace boost::get with mapbox::util::get
Signed-off-by: James Ward <james.ward@arm.com>
Change-Id: I38460cabbcd5e56d4d61151bfe3dcb5681ce696e
Jan Eilers [Wed, 9 Sep 2020 18:11:16 +0000 (19:11 +0100)]
IVGCVSW-5295 Remove boost::format from armnn profiling
* Replace all instances of boost::format with fmt::format from armnn/src/profiling
Signed-off-by: Jan Eilers <jan.eilers@arm.com>
Change-Id: I4722bdde52c740bc9bdce969128074a34a3ee75e
Mike Kelly [Tue, 29 Sep 2020 19:50:35 +0000 (20:50 +0100)]
Refactored Optimize(...) function to throw exceptions instead of returning null
* INetwork::Optimize(...) states that the function should throw an exception
if it fails but the implementation in Network.cpp returned null in some
scenarios instead. This has led to some confusion amongst users.
Signed-off-by: Mike Kelly <mike.kelly@arm.com>
Change-Id: I358d1293232c9464772aa0e39ab3355e3570c823
Teresa Charlin [Tue, 29 Sep 2020 12:27:07 +0000 (13:27 +0100)]
Update ACL pin to
d6d1b3682a2cdd54bae5498635b108a4b19a045a
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: I86f5848e895536334e3c6353e721ced871a192e0
Derek Lamberti [Mon, 28 Sep 2020 15:11:50 +0000 (16:11 +0100)]
Quantization copy constructor
* Fix compiler implicit copy deprecation warning.
* Simplify copy operator by removing unnecessary equality comparison.
Now just does pointer comparison instead.
Change-Id: I9ebe170c637c636919b9d8729a78449148b7f83e
Signed-off-by: Derek Lamberti <derek.lamberti@arm.com>
James Ward [Fri, 25 Sep 2020 10:43:21 +0000 (11:43 +0100)]
IVGCVSW-4519 Remove Boost Variant and apply_visitor variant
* add mapbox/variant third party package
Signed-off-by: James Ward <james.ward@arm.com>
Change-Id: I181302780edd9dace40f158a11327316a12ce69a
Jan Eilers [Fri, 25 Sep 2020 07:36:44 +0000 (08:36 +0100)]
IVGCVSW-5295 Change fmt to be a header-only interface library
* Fix non-virtual-dtor warnings in fmt
* Fix wrong fmt include in TfParser
* Make fmt work in nn-driver
* Make fmt a header-only interface library
* Link fmt where necessary
Signed-off-by: Jan Eilers <jan.eilers@arm.com>
Change-Id: I9db7cd9a133a81262cbf79f04fb419ab97b88ea8
Nikhil Raj [Tue, 29 Sep 2020 13:34:34 +0000 (14:34 +0100)]
Update ACL pin to
d175ecea20f5c9e3bff07f7c5d3e4bb5519f7cb9
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: Ib0b28b3840eb8c8d38ad32e1a572cd9a75ff8a82
Matthew Sloyan [Fri, 11 Sep 2020 09:14:57 +0000 (10:14 +0100)]
IVGCVSW-5301 Remove all boost::numeric_cast from armnn/src/profiling
* Replaced with armnn/utility/NumericCast.hpp
Signed-off-by: Matthew Sloyan <matthew.sloyan@arm.com>
Change-Id: I14e89c00cd1dd607315113d5b8ca56c2e9f6c363
Francis Murtagh [Mon, 21 Sep 2020 14:27:00 +0000 (15:27 +0100)]
IVGCVSW-5326 PyArmnn has hard dependencies on all parsers when using cmake
Signed-off-by: Francis Murtagh <francis.murtagh@arm.com>
Change-Id: Ieffc1f3db548bc01ea0127a0d60659f0b0783ce3
Finn Williams [Thu, 17 Sep 2020 14:58:31 +0000 (15:58 +0100)]
IVGCVSW-5325 Speed up the reference backend
Change-Id: Id8bd0a0418be31d975b944b54bbacb25051ffb2e
Signed-off-by: Finn Williams <Finn.Williams@arm.com>
Teresa Charlin [Fri, 25 Sep 2020 14:08:21 +0000 (15:08 +0100)]
IVGCVSW-4973 Enable QLstm projection unit tests on CL
*Cosmetic changes on ClQLstmWorkload
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: I61f55343263e623aaae042d8dfe8c294540e98f1
Nikhil Raj [Thu, 24 Sep 2020 16:58:34 +0000 (17:58 +0100)]
Update ACL pin to
840a72cc745c60eccbd26fe192b035ec68b2ee41
* Change tensor to non const to fix build error caused by ACL fix for QLSTM
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: I7ab0f644dfb3cb3cf21bda73028e9368f3354f4a
Inki Dae [Thu, 10 Sep 2020 06:33:54 +0000 (15:33 +0900)]
Add int32 and int64 ArgMax op support
This patch adds int32 and int64 ArgMax op support.
Current ARMNN already has ArgMax op but not used, and
it doesn't support int64 output type.
So this patch adds a new type, Signed64, and also adds
ArgMinMax computation function for int64 type support.
In default, output tensor type of ArgMax op is int64 in case of
tensorflow lite model so this patch makes a proper function - ArgMax op
for int64 or int32 - to be called according to parsed output_type value.
With this patch, ARMNN supports both types - int64 and int32 - for
ArgMinMax op.
Changelog v1:
- Check if output data type of ArgMinMax op is valid or not.
- Use template function to support int32 and int64 types of ArgMinMax function.
- Keep using Signed32 as default data type of m_Output_Type.
Change-Id: I7a8e7e38dd9e5acc81464571d8b4d51378fc7f14
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Nikhil Raj [Wed, 23 Sep 2020 11:50:08 +0000 (12:50 +0100)]
Update ACL pint to
1643a45557fde79ee209f55c507860307ffe627c
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: Id8e240a711064c48e4bdb9b1236198647e365a7f
Sadik Armagan [Tue, 22 Sep 2020 13:35:19 +0000 (14:35 +0100)]
IVGCVSW-5318 'Create a Neon/CL Workload Unit Test fast_math option enabled'
* Unit test implemented to make sure it returns WINOGRAD
* Updated the enable-fast-math option in ExecuteNetwork to be consistent
Signed-off-by: Sadik Armagan <sadik.armagan@arm.com>
Change-Id: Id64f114ae47966def69a9eef0770a4251ee56a41
Nikhil Raj [Tue, 22 Sep 2020 12:07:49 +0000 (13:07 +0100)]
Update ACL pin to
34654b2d8dcaf268a9d1bf9e0cdb5ba548ced2b7
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: I3601db66ea270d6799627ac96ec95a9008d6942f
Mike Kelly [Tue, 15 Sep 2020 12:37:38 +0000 (13:37 +0100)]
Added Security Issues section
* Added Security Issues section with email address.
Signed-off-by: Mike Kelly <mike.kelly@arm.com>
Change-Id: If1ccfd3c7847cac0a09b6ec72fdfacba0011455e
Matthew Sloyan [Mon, 14 Sep 2020 10:47:26 +0000 (11:47 +0100)]
IVGCVSW-5303 Remove some boost::numeric_cast from rest of ArmNN
* Replaced with armnn/utility/NumericCast.hpp
* Exclusions in TypeUtils.cpp and QuantizerVisitor.cpp
* Excluded as requires float implementation in NumericCast.hpp
Signed-off-by: Matthew Sloyan <matthew.sloyan@arm.com>
Change-Id: I5c4c60e7028e1a51bf9379457278d253fd37bc70
Matthew Sloyan [Wed, 9 Sep 2020 08:07:37 +0000 (09:07 +0100)]
IVGCVSW-5300 Remove some boost::numeric_cast from armnn/backends
* Replaced with armnn/utility/NumericCast.hpp
* Some exclusions in reference backend
* Excluded as requires float implementation in NumericCast.hpp
Signed-off-by: Matthew Sloyan <matthew.sloyan@arm.com>
Change-Id: I9e4e9cd502c865452128fa04415fd6f250baa855
Sadik Armagan [Tue, 15 Sep 2020 16:17:08 +0000 (17:17 +0100)]
IVGCVSW-5317 'Add enable_fast_math Option to ExecuteNetwork'
Signed-off-by: Sadik Armagan <sadik.armagan@arm.com>
Change-Id: I4eb3e27837aea926593d49f9ccea07bab8388d5b
Nikhil Raj [Fri, 11 Sep 2020 09:19:42 +0000 (10:19 +0100)]
Update ACL pin to
16cf3c2318978580006633ea4d5eccdd8ba1194a
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: I6f1f66fdd2e49ca36d28fd2367f4b64dc6d38cca
Narumol Prangnawarat [Mon, 14 Sep 2020 15:12:44 +0000 (16:12 +0100)]
IVGCVSW-5305 AddBroadcastReshapeLayer as optimizer
* Remove AddBroadcastReshapeLayer from TfLiteParser
* Add AddBroadcastReshapeLayer as optimizer
* AddBroadcastReshapeLayer optimizer unit tests
* Load-scope dynamic tensor broadcasting unit tests
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
Change-Id: I3549e85b71b41cbd4d96c0f1ece7887acbca76d1
Éanna Ó Catháin [Mon, 14 Sep 2020 16:36:49 +0000 (17:36 +0100)]
MLECO-929 Add Object Detection sample application using the public ArmNN C++ API
Change-Id: I14aa1b4b726212cffbefd6687203f93f936fa872
Signed-off-by: Éanna Ó Catháin <eanna.ocathain@arm.com>
Matthew Sloyan [Fri, 11 Sep 2020 15:17:48 +0000 (16:17 +0100)]
IVGCVSW-5302 Remove some boost::numeric_cast from parsers
* Replaced with armnn/utility/NumericCast.hpp
* Exclusions in armnnCaffeParser
* Three excluded as requires float implementation in NumericCast.hpp
Signed-off-by: Matthew Sloyan <matthew.sloyan@arm.com>
Change-Id: Ib468b606238694334a8319d0ed5db381ce37a915
Sadik Armagan [Mon, 14 Sep 2020 14:44:18 +0000 (15:44 +0100)]
IVGCVSW-5157 'Pipe ModelOption through Network::LoadNetwork() to Workload factory'
* Pass ModelOptions to WorkloadFactory
* Updated signature of CL and NEON Convolution2d workloads added FastMathEnabled param.
Signed-off-by: Sadik Armagan <sadik.armagan@arm.com>
Change-Id: I536178be8e4dd4083489e69febadaf0feeba46d2
Matthew Sloyan [Tue, 8 Sep 2020 11:00:32 +0000 (12:00 +0100)]
IVGCVSW-5299 Remove some boost::numeric_cast from armnn/tests
* Replaced with armnn/utility/NumericCast.hpp
* Removed combinations without float implementation in NumericCast.hpp
Signed-off-by: Matthew Sloyan <matthew.sloyan@arm.com>
Change-Id: Ia4ec605f063cdb0071fff302ef48c610f9f9505e
Nikhil Raj [Thu, 10 Sep 2020 09:21:00 +0000 (10:21 +0100)]
Update ACL pin to
5489394cdfe10afb469171cf521f26f923eeb2e2
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: I05375f653f76084c64df47390acf1ec692cb16b6
Kevin May [Tue, 8 Sep 2020 14:50:18 +0000 (15:50 +0100)]
IVGCVSW-5245 Add Quantization operator=() function
* Add unit tests to check if Quantization info copied correctly
Signed-off-by: Kevin May <kevin.may@arm.com>
Change-Id: I7bb7bde5d97e82c57252c6d5131fbe21ad3096d2
Sadik Armagan [Thu, 10 Sep 2020 12:37:32 +0000 (13:37 +0100)]
IVGCVSW-5156 Introduce ModelOptions to OptimizedNetwork
* Introduced ModelOptions to IBackendInternal
* Introduced ModelOptions to Network
* Added FastMathEnabled parameter to Conv2d Validate function in CL and NEON
* Added Optimizer tests
Signed-off-by: Ryan OShea <Ryan.OShea2@arm.com>
Signed-off-by: Sadik Armagan <sadik.armagan@arm.com>
Change-Id: Ib54c1e82cb3d89a52756ed499cf91b6a7fdb2063
James Ward [Mon, 7 Sep 2020 15:45:07 +0000 (16:45 +0100)]
IVGCVSW-5293 Remove boost::format from armnn/tests
* Replaced boost::format with fmt::format
Signed-off-by: James Ward <james.ward@arm.com>
Change-Id: Icf5a6508e7be3d31bc063643491fc5e0607f21fa
Jan Eilers [Tue, 8 Sep 2020 07:57:40 +0000 (08:57 +0100)]
IVGCVSW-5197 Add support for 2nd input to ExpandDims of TfParser
* ParseExpandDims did not support to pass the axis parameter as
a second input tensor
* Added related unit tests
Signed-off-by: Jan Eilers <jan.eilers@arm.com>
Change-Id: I8217950f0b42beaf5b9eaebdcad04267e4443ba3
Vincent ABRIOU [Thu, 3 Sep 2020 11:02:43 +0000 (13:02 +0200)]
fix cxxopts and ghc cross compilation issue
While cross compiling using openembedded yocto environment,
CXXOPTS_INCLUDE and GHC_INCLUDE variables remains empty because the search
path point to the openembedded sysroot instead of the current directory.
To avoid this situation, NO_CMAKE_FIND_ROOT_PATH must be added to the
find_path call.
Signed-off-by: Vincent ABRIOU <vincent.abriou@st.com>
Change-Id: I3ba17a300d5ca7fbc58b6ec5c4eb15267dcee228
Nikhil Raj [Wed, 9 Sep 2020 10:41:19 +0000 (11:41 +0100)]
Update ACL pin to
57f30a9309ff2e5e3b32731a785bf38b01d1fd69
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: Ie34332830e52e9c5aa35e8641ffdf562f33e6b93