Age | Commit message (Collapse) | Author |
|
- Remove Bf16ToFp32 Conversion Layer
- Remove Fp32ToBf16 Conversion Layer
- Remove B16 Conversion tests
* Throw exception if m_ReduceFp32ToBf16 optimzer option is set to true
* Provide comments to enable fast math in order to use bf16
* Update docs to inform users to enable fast math for bf16
Execute Network Changes
* Require bf16_turbo_mode to also have fast_math_enabled set to true
- Remove setting m_ReduceFp32ToBf16 optimizer option
Signed-off-by: Ryan OShea <ryan.oshea3@arm.com>
Change-Id: Ibaa6da9d29c96a1ce32ff5196b0847fde9f04a1c
|
|
* Add functionality to print output tensors to file in tempdir
* UnitTests
Signed-off-by: Keith Davis <keith.davis@arm.com>
Change-Id: Idfb4c186544187db1fecdfca11c662540f645439
|
|
Fp32NetworkToBf16Converter
* Fuse FP32ToBF16Layers with Constant Layer so Conv2d/FullyConnected
can have their weights redirected.
* If BF16 Unsupported in Conv2d || FullyConnected revert fused
Constant Layer to FP32
Change-Id: If523c708a822659d64597d9ae39cca1c2f84b76f
Signed-off-by: Francis Murtagh <francis.murtagh@arm.com>
|
|
* Update Front-end and Tools.
* Updated Serializer, Deserializer and unit tests to reflect this.
* Updated TfLiteDelegate, TfLiteParser and OnnxParser.
* Updated Ref.
* Fixed resulting Neon / CL tests
* Unified optimizers for conv2d ops
* Optimizer Fix - Fp32ToBf16
* Partial implementation for ACL backends to fix VTS failures
!android-nn-driver:7477
Signed-off-by: Keith Davis <keith.davis@arm.com>
Change-Id: I5fb18877f7ee32643e15a9818945356274bb401b
|
|
Debug layers are given names based on the layer they take their input from.
If a layer has multiple outputs then one debug layer will be attached to
each output. Therefore all these debug layers would have identical names
making them hard to distinguish when debugging. This patch includes the
output slot index which the debug layer takes its input from when creating
the name.
Change-Id: I09eaa8a7edad9bfdf678b4778cf740340013126c
Signed-off-by: Rob Hughes <robert.hughes@arm.com>
|
|
* Change boost assert to armnn assert
* Change include file to armnn assert
* Fix ARMNN_ASSERT_MSG issue with multiple conditions
* Change BOOST_ASSERT to BOOST_TEST where appropriate
* Remove unused include statements
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
Change-Id: I5d0fa3a37b7c1c921216de68f0073aa34702c9ff
|
|
Convolution2d and FullyConnected layers
* Add InsertConvertFp32ToBf16LayersBefore
* Add ConvertWeight to ConvertFp32NetworkToBf16Impl for Conv2d and FullyConnected
* Allow different input and output when input is BF16 and output is FP32
Conv2d and FullyConnected layers
* Unit tests
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
Change-Id: Ic8f92ff28edcae08a72a3114a28f50c4619f919b
|
|
* Add ReduceFp32ToBf16 to OptimizerOptions
* Add ConvertFp32NetworkToBf16
* Add utility functions to insert conversion layers
* Add constant conversion BF16 <-> FP32
* Unit tests
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
Change-Id: Iaff77e20c721400b052cb37eb9ef6fe16d7abaff
|
|
Dequantize
* Check for output data type as well as input data type when determining
whether we should attempt to fall back to FP32 if FP16 is not supported
* Override output type for Dequantize in IsLayerSupported() instead of
input type
* Updated original input type from FP16 to FP32 in InsertConvertFp32ToFp16LayersAfter()
Signed-off-by: Aron Virginas-Tar <Aron.Virginas-Tar@arm.com>
Change-Id: Ic6477fd17cea5a91bd8bf9ae0cf836520897d5b7
|
|
* Moved to BackendRegistry.hpp include/armnn
* Updated makefiles and sources accordingly
Signed-off-by: Matteo Martincigh <matteo.martincigh@arm.com>
Change-Id: I4d83abb581d523218a880c879fcf30c9611f7fd7
|
|
Take a copy of the MemoryStrategies for a layer before inserting
new connections. Use the copy when looking up the original MemoryStrategies
during the graph transformation.
Fix the unit tests for AddCopyLayers to have cases where copies are needed.
Fix the validation for clarity and correctness - was previously comparing
Layers by pointer when it should have been by name (as it was comparing
with a cloned graph).
Change-Id: Ie282dc11913e977b8151ce1ad8bfba5e11617d40
Signed-off-by: Matthew Bentham <Matthew.Bentham@arm.com>
|
|
Change-Id: Ie50aeccf053c20c3a01a75042bbc3acd824375af
Signed-off-by: Derek Lamberti <derek.lamberti@arm.com>
Signed-off-by: Matteo Martincigh <matteo.martincigh@arm.com>
|
|
* Also update Debug layer to use layer guid information
Change-Id: I9ec1f639299c3f855b670ff031a0e88d685cfc6b
Signed-off-by: Nattapat Chaimanowong <nattapat.chaimanowong@arm.com>
|
|
Change-Id: Idda9c8ba244c74d931105ced90c38a12d48df48a
Signed-off-by: Aron Virginas-Tar <Aron.Virginas-Tar@arm.com>
|
|
subgraphs are optimized by the backends
* Added a new method OptimizeSubGraph to the backend interface
* Refactored the Optimize function so that the backend-specific
optimization is performed by the backend itself (through the new
OptimizeSubGraph interface method)
* Added a new ApplyBackendOptimizations function to apply the new
changes
* Added some new convenient constructors to the SubGraph class
* Added AddLayer method and a pointer to the parent graph to the
SubGraph class
* Updated the sub-graph unit tests to match the changes
* Added SelectSubGraphs and ReplaceSubGraphConnections overloads
that work with sub-graphs
* Removed unused code and minor refactoring where necessary
Change-Id: I46181794c6a9e3b10558944f804e06a8f693a6d0
|
|
separate branch) and master
* Brings in all the changes done for the pluggable backends
* Added sub-graph support and tests
* Added precompiled layer support and tests
* Moved BackendSettings to a separate file
* Removed the backend-specific code
* Ported DebugLayer and associated functionality
* Included fixes to make those changes work with master
Change-Id: Id7028fa7917527b844628d5aff5732e3d94c0488
|
|
* Modified optimizer to support debug mode via DebugLayer
Change-Id: Ic8f313778e55540c182cf99876c44a0823be04c6
|
|
Change-Id: I26fb13717f64c942b2648702c52be18be856042d
|