Age | Commit message (Collapse) | Author |
|
* TfLite slice parser and relevant tests added
* TfLite transpose parser logic added to translate Tf/np permutation
vector definitions to Armnn definitions
* TfLite transpose parser no permute data test modified to include
data for default permutation vector when none specified
Signed-off-by: josh minor <josh.minor@arm.com>
Change-Id: Iebd30971bd180593dc6b8f0d5be1d1bc61a3a5bf
|
|
* Moved to BackendRegistry.hpp include/armnn
* Updated makefiles and sources accordingly
Signed-off-by: Matteo Martincigh <matteo.martincigh@arm.com>
Change-Id: I4d83abb581d523218a880c879fcf30c9611f7fd7
|
|
* Moved QuntizeHelper.hpp to armnnUtils
* Reordered parameters for QuantizedVector and added default
values for qScale and qOffset to make life easier when
using the function for non-quantized types such as Float16
Signed-off-by: Aron Virginas-Tar <Aron.Virginas-Tar@arm.com>
Change-Id: I28c263dfa425f1316feccb4116839a84f5d568e5
|
|
* Fixed bug in custom operator support that caused all custom operators
to be parsed as a DetectionPostProcessLayer
* Added support for handling unsupported operators (built-in or custom)
by replacing them with a StandInLayer in the generated network
* Added options to TfLiteParser to control whether we want to use
StandInLayers when we encounter unsupported operators, or we prefer
to throw a ParserException as until now
Signed-off-by: Aron Virginas-Tar <Aron.Virginas-Tar@arm.com>
Change-Id: I125a63016c7c510b1fdde6033842db4f276718c4
|
|
Signed-off-by: Francis Murtagh <francis.murtagh@arm.com>
Change-Id: I6c69f1f2f0f9c49ddada57b41d5802919ea14b25
|
|
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
Change-Id: I026e4880013fa4ed83b6c4643dda7e4d100014a5
|
|
* Improved parsing of the ArmNN's version at compile time
* Reading the version directly from the Version.hpp file, to
make sure not to break any build made with a different
build system than CMake (or Ninja)
* The version macro is set in the Version.hpp just like before,
so setting it from the command line is no longer necessary
!android-nn-driver:2126
Signed-off-by: Guillaume Gardet <guillaume.gardet@arm.com>
Signed-off-by: Matteo Martincigh <matteo.martincigh@arm.com>
Change-Id: I0820094607734d283ded030e4ded0ae6f5889c86
|
|
* Add permute vector to descriptor if present
* Refactor test to check with and without permute vector
Signed-off-by: Kevin May <kevin.may@arm.com>
Change-Id: Ic8d882bb0f982fd00bb2854c18ea316b1b2cde2b
|
|
Signed-off-by: Aron Virginas-Tar <Aron.Virginas-Tar@arm.com>
Change-Id: I766f4297b9daa26edacc2079fe62a083ba2fa68f
|
|
Signed-off-by: Keith Davis <keith.davis@arm.com>
Change-Id: I94e975ede3876c812f0f26b040e8c50a5597f86a
|
|
* FCRN and FSRCNN aren't supported on all backends yet.
Signed-off-by: Mike Kelly <mike.kelly@arm.com>
Change-Id: Ibaddfba731224aac24de05c225fb94efd3fe7f94
|
|
* CMake "install" commands require a RUNTIME argument for platforms with
DLLs (e.g. Windows).
* Replace use of non-standard variable length array with vector
* Remove unnecessary #include of unistd.h
* Add #ifdefs to dynamic backend code to disable for non-Unix platforms
where you can't use dlopen etc. We could implement this properly for Windows
later using LoadLibrary etc., but for now erroring is fine.
* Add missing #include of <algorithm>
Change-Id: Ic8ef5fd599b37bf8772510157b6e479819f6a1eb
|
|
* Move DeepSpeaker, FCRN and FSRCNN from TensorFlowSupport.md
to TensorFlowLiteSupport.md
Signed-off-by: Aron Virginas-Tar <Aron.Virginas-Tar@arm.com>
Change-Id: I82d508d3599073eb8f6638994a91f46b4b95c247
|
|
Signed-off-by: Aron Virginas-Tar <Aron.Virginas-Tar@arm.com>
Change-Id: I3da43122240fda7864e0932d47b3efeea8bbed40
|
|
* Reordered and edited inputs in ParseTransposeConv method
* Updated test cases to reflect changes
Signed-off-by: Matthew Jackson <matthew.jackson@arm.com>
Change-Id: Icd090cf6ab340eed4f098ef8641daf7c6e0d0949
|
|
* Added ParseTransposeConv to TfLite Parser
* New TransposeConv test file
* Updated documentation for supported Transpose Convolution
Change-Id: Id7356d8525556805c164af693ae2b16f6a8492fa
Signed-off-by: Matthew Jackson <matthew.jackson@arm.com>
|
|
* Increased MaxNumOfTensorDimensions and fixed issues related to its use
* Fixed issues caused by assuming 5d tensors are invalid
* Updated ArmComputeTensorUtils for 5d tensors
* Added 5d tensor unit tests for add, mul, stack and reshape (needed by IVGCVSW-3527)
Signed-off-by: Matthew Jackson <matthew.jackson@arm.com>
Change-Id: I5bcd64942d0d04efcc6c5acb240ad4b88e010743
|
|
* experimental/filesystem is unavailable in some versions of LLVM
Change-Id: I0136bc911f2607561e8000fe9fa7c07f480554b6
Signed-off-by: Mike Kelly <mike.kelly@arm.com>
|
|
* Remove multiplier check in TfLite parser
* Add reference unit test for depthwise multipler of 64, as in DeepSpeaker
Signed-off-by: Matthew Jackson <matthew.jackson@arm.com>
Change-Id: I787339ab4c4d269333985353d191202d070906ba
|
|
* Added ParseL2Normalization in TfLiteParser
* Added new unit tests L2Normalization.cpp
* Added documentation for supported L2 Normalization to TensorflorLiteSupport.md
Signed-off-by: Matthew Jackson <matthew.jackson@arm.com>
Change-Id: I83ea75d1791ac8a00390aed3e5d0a7b337fcd46d
|
|
* Added ParsePack method
* New unit test Pack.cpp
* Updated TensorFlowLiteSupport.md with new supported operator
Signed-off-by: Matthew Jackson <matthew.jackson@arm.com>
Change-Id: I2310b33ee26959b036bb4452a25c90cc1d4cbf20
|
|
!android-nn-driver:1451
Signed-off-by: Aron Virginas-Tar <Aron.Virginas-Tar@arm.com>
Change-Id: Ieedbce1f6e95891137a250fdd07e2f7e4e1f4828
|
|
Change-Id: If38336f1678504849edb0134a59daae1c8d9ef92
Signed-off-by: Matthew Bentham <matthew.bentham@arm.com>
|
|
We had check in the parser to catch calls where deph_mult > 1, this came from
the time when ACL supported only multiplier == 1.
Change-Id: Ic21a2cd6a5bb328d43fc9c667e847429a57760b2
Signed-off-by: Pablo Tello <pablo.tello@arm.com>
|
|
* Add check if Sub-tensors cannot be used, call ACL function
* Add computation of SplitAxis from SplitterDescriptor
* Add NeonSplitterWorkload functions
* Modify IsSplitterSupported to call ACL validate function
if sub-tensor cannot be used
* Also check if quantization parameters match when using sub-tensors
* Add more unit tests for Splitter in TfParser and TfLiteParser
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
Change-Id: I31e4c7d055117c83c65b598c4125442173242226
|
|
* Used the new ARMNN_DEPRECATED_MSG macro instead of @deprecated
* Refactored the code to no longer use the deprecated methods where
applicable
!android-nn-driver:1126
Change-Id: Ib0578d3d6fc5a763f5fb922f67ba91fafc7796f6
Signed-off-by: Matteo Martincigh <matteo.martincigh@arm.com>
|
|
!android-nn-driver:1120
Change-Id: I5192fa3deb4ea9766d38ad0bf4dfbfa0b4924c41
Signed-off-by: Jim Flynn <jim.flynn@arm.com>
|
|
Added support for dilation in DepthwiseConvolution2d in the
Neon and CL backends.
Change-Id: Ie1522b498c07f80d6efcf9dc79e926c8cfa06ca5
Signed-off-by: Pablo Tello <pablo.tello@arm.com>
|
|
Change-Id: I3deb2b9a37e8a8f8f2ed93c64ed0656ae911e24c
Signed-off-by: Jim Flynn <jim.flynn@arm.com>
|
|
* Get outputTensorInfo from runtime outputs to ensure that
the results are from the running network
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
Change-Id: Ib35afeaf5f6121df7f6379cfc557ae77e5214d69
|
|
Change-Id: Ie50aeccf053c20c3a01a75042bbc3acd824375af
Signed-off-by: Derek Lamberti <derek.lamberti@arm.com>
Signed-off-by: Matteo Martincigh <matteo.martincigh@arm.com>
|
|
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
Change-Id: I18697fde21e90723bcc36cc9cc1b571652fa54e7
|
|
when input is > 2D to flatten the input to 2D [batch_size, input_size]
Change-Id: Id9d9ff996225c7d0938204ae0ceb330a11e264f5
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
|
|
* Added a check after enqueue workload to ensure that the outputs have the correct number of dimensions
* OutputTensors cannot be used for this as in RunTest we're specifically creating these with expected number of dimensions
Signed-off-by: Nina Drozd <nina.drozd@arm.com>
Change-Id: Ib6e5a138240e2f639f462f58caa72ae760e2b406
|
|
to remove the unpacked dimension of each output from Splitter
and correct ReshapeFixtureWithReshapeDimsFlatten test output shape
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
Change-Id: I517d315475612ac8b773930f9b58cac316fa8553
|
|
This line is definitely wrong, plenty of float32 things
work in the TFLite parser, so it sets the wrong expectations.
Change-Id: I87ac01e14b916ccd767429c124c353ab136174d8
Signed-off-by: Matthew Bentham <matthew.bentham@arm.com>
|
|
* Allow input data with dimension not greater than 4D
* Correct input order
* Get split dimension from buffer data
* Unit tests
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
Change-Id: I285851b19e6fa7c715e5fe4853df167e7c856647
|
|
Change-Id: I6dab12aed395a30466d66421c6e5a12659fedac8
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
|
|
* Added ParseUnpack in TfLiteParser
* New Unpack test file with test reproducing unpack in DeepSpeechV1 model
* Added documentation for supported Unpack to TensorflorLiteSupport.md
Signed-off-by: Nina Drozd <nina.drozd@arm.com>
Change-Id: Ie920d46254ff4b4ab544407ace4c1d489af83157
|
|
!android-nn-driver:968
Signed-off-by: Aron Virginas-Tar <Aron.Virginas-Tar@arm.com>
Change-Id: I03ccb4842b060a9893567542bfcadc180bbc7311
|
|
* Added ParseTanH in TfLiteParser
* Added testcase for parsing TanH activation in Activations.cpp
* Added new supported TanH operator in TensorflowLiteSupport.md
Signed-off-by: Nina Drozd <nina.drozd@arm.com>
Change-Id: Ie60f38994d76b077f86828c21c57381bdeeaeb5a
|
|
* Added ParseSplit method
* New Unit test Split.cpp
* Updated TensorflowLiteSupport.md with new supported operator
Change-Id: Iec80ba9ad7b48db8e86589ebae77bd7d8ed38fb2
Signed-off-by: Nina Drozd <nina.drozd@arm.com>
|
|
* Add Parse Exception for convolutions without default dilation
Signed-off-by: Kevin May <kevin.may@arm.com>
Change-Id: I1b8f75c2d871d81161eb5378ced277438e809ba2
|
|
Change-Id: I06653135c5fef38a52995da6c4b6de7ba5786b6a
Signed-off-by: Matthew Bentham <matthew.bentham@arm.com>
|
|
* Add helper function to generate custom data for detectPostProcess
* Test helper function within current test suite
Change-Id: I9e66d0a28d69b1376da67723f03b112d17e97281
Signed-off-by: keidav01 <keith.davis@arm.com>
Signed-off-by: Aron Virginas-Tar <Aron.Virginas-Tar@arm.com>
|
|
!referencetests:176231
Change-Id: Ic8a641507c6af2f38e535b0208eb406e74a334a1
Signed-off-by: Ruomei Yan <ruomei.yan@arm.com>
|
|
* Add support for Custom operator - TFLite_Detection_PostProcess
* Add Quantized SSD MobileNet to tested networks
Change-Id: Id7e729ed02dbc401638b9832c8b9b5d71a6d31b7
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
|
|
* Assign output shape of MobileNet SSD to ArmNN network
* Add m_OverridenOutputShapes to TfLiteParser to set shape in GetNetworkOutputBindingInfo
* Use input quantization instead of output quantization params
* Correct data and datatype in Inference test
Change-Id: I01ac2e07ed08e8928ba0df33a4847399e1dd8394
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
Signed-off-by: Aron Virginas-Tar <Aron.Virginas-Tar@arm.com>
|
|
* Added additional custom option use_regular_nms to JSON string
* Renamed use_regular_non_max_suppression to use_regular_nms in parser
Change-Id: I5431eabade6337f6f8c3372670862e32e9cda63c
Signed-off-by: keidav01 <keith.davis@arm.com>
|
|
Change-Id: Ia349d71ec966b37feaffb971c7a77bf8cd44501c
Signed-off-by: Derek Lamberti <derek.lamberti@arm.com>
|