aboutsummaryrefslogtreecommitdiff
path: root/src/backends/reference/workloads/Decoders.hpp
AgeCommit message (Collapse)Author
2022-11-16IVGCVSW-7214 Disable BF16-Turbo-Mode and remove conversion layersRyan OShea
- 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
2021-10-02IVGCVSW-5985 Remove deprecated codeJan Eilers
* Removes deprecated AddLayer, IsLayerSupported functions * Marks the whole LayerVisitor class as deprecated not just the constructor. This required to wrap all Accept functions in a no deprecate macro because the LayerVisitor is used as a parameter in there * Removes usage of deprecated LayerVisitor and replaces it with ExecuteStrategy. This required a few structural changes in the unit tests * Adds a default implementation for IStrategy called StrategyBase * Changes pyarmnn to use non deprecated constructor for INetworkProperties and adds related unit test * Marks usage of deprecated code in pyarmnn as deprecated. This required to extend INetworkProperties to allow backwards compatibility * Removes deprecated functions from CpuAcc, GpuAcc and Ref backends Note: This patch breaks compatibility with backends that are not updated in this patch !android-nn-driver:6325 Signed-off-by: Jan Eilers <jan.eilers@arm.com> Change-Id: Id13b6f37a74d26eadeda2da1dc92915e725ed5a5
2021-06-16IVGCVSW-5826 Change weights layout for depthwise to [1,H,W,I*M]Jan Eilers
* This change is necessary because tflite uses a [1,H,W,I*M] format and uses the I*M dimension for per axis quantization. Our previous layout [M,I,H,W] can't handle the correlating quantization scales. * Updates Onnx-, TfLiteParser and TfliteDelegate * Updates the CpuRef, CpuAcc and GpuAcc backends * Adjusts unit tests * Adds test to ensure models with old layout can still be read and executed * Adds conversion function to previous layout [1,H,W,I*M] --> [M,I,H,W] which can be used by backend developers !android-nn-driver:5553 Signed-off-by: Jan Eilers <jan.eilers@arm.com> Change-Id: Ifef23368b8c3702cf315a5838d214f7dc13c0152
2020-11-09IVGCVSW-5091 Add Logical ops frontend and ref implJames Conroy
* Add frontend and reference implementation for logical ops NOT, AND, OR. * Unary NOT uses existing ElementwiseUnary layer and ElementwiseUnary descriptor. * Binary AND/OR uses new layer LogicalBinary and new LogicalBinary descriptor. * Add serialization/deserializion support and add missing ElementwiseUnary deserializer code. * Add additional Boolean decoder in BaseIterator.hpp. Signed-off-by: James Conroy <james.conroy@arm.com> Change-Id: Id343b01174053a166de1b98b6175e04a5065f720
2020-06-30IVGCVSW-5007 Implement an Int32 reference Elementwise workloadFinn Williams
Signed-off-by: Finn Williams <Finn.Williams@arm.com> Change-Id: I6592169b74ac4294bc09647879aec0718c641f91
2020-04-06IVGCVSW-4485 Remove Boost assertNarumol Prangnawarat
* 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
2020-03-19IVGCVSW-4565 TENSOR_BOOL8 data type not supported in AndroidNN DriverSadik Armagan
* Enabled Boolean and Int32 data types in Reference Comparison inputs * Added decoder for Boolean data type * Refactored ClGreaterWorkload to work with any data type * Refactored NeonGreaterWorkload to work with any data type !android-nn-driver:2902 Signed-off-by: Sadik Armagan <sadik.armagan@arm.com> Change-Id: I92772810b744b388831c9dca0119ebf8cb7a447e
2020-03-09IVGCVSW-4517 Implement BFloat16 Encoder and DecoderNarumol Prangnawarat
* Add ConvertFloat32ToBFloat16 * Add ConvertBFloat16ToFloat32 * Add BFloat16Encoder * Add BFloat16Decoder * Unit tests Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com> Change-Id: I198888384c923aba28cfbed09a02edc6f8194b3e
2020-02-07IVGCVSW-4386 Add ArmNN reference support for QAsymmS8Ryan OShea
* Added Quantization Scheme for QAsymmS8 * Added Unit Tests for QAsymmS8 * Renamed QAsymm8 calls to QAsymmU8 Signed-off-by: Ryan OShea <Ryan.OShea2@arm.com> Change-Id: I897b4e018ba1d808cc3f8c113f2be2dbad49c8db
2020-01-24IVGCVSW-4370 Deprecate DataType::QuantizedSymm8PerAxisDerek Lamberti
!android-nn-driver:2622 Change-Id: If99d3eff71ff66ba28af1e5af248299fe04511b9 Signed-off-by: Derek Lamberti <derek.lamberti@arm.com>
2020-01-13Rename quantized data types to remove ambiguity for signed/unsigned payloadsDerek Lamberti
!android-nn-driver:2572 Change-Id: I8fe52ceb09987b3d05c539409510f535165455cc Signed-off-by: Derek Lamberti <derek.lamberti@arm.com>
2019-12-09IVGCVSW-4211 Add Signed 8 bit Quantisation support into the Reference backendFinn Williams
!android-nn-driver:2435 Signed-off-by: Finn Williams <Finn.Williams@arm.com> Change-Id: I10ecd4a8937725953396805f33a3562a5384c4d4
2019-11-29IVGCVSW-4209 Create a public API for the ArmNN UtilsMatteo Martincigh
* Moved the relevant armnnUtils headers to the new location: include/armnnUtils * Update the header usage throughout the source code !android-nn-driver:2387 Signed-off-by: Matteo Martincigh <matteo.martincigh@arm.com> Change-Id: I2ba15cebcacafad2b5a1a7b9c3312ffc585e09d6
2019-11-05IVGCVSW-3836 Add support for Int32 per-axis scalesAron Virginas-Tar
* Added ScaledInt32PerAxisDecoder implementation * Added new case for Signed32 in MakeDecoder that returns a ScaledInt32PerAxisDecoder if the tensor info has multiple quantization scales Signed-off-by: Aron Virginas-Tar <Aron.Virginas-Tar@arm.com> Change-Id: I8b3c11091644da993044d2a0fe2aba6b06b5af56
2019-11-04IVGCVSW-3835 Create Encoder and Decoder for QSymm8PerAxisKeith Davis
* Add QuantizedSymm8PerAxis to armnn DataType (types.hpp) and * Add Quantize and Dequantize template for int8 in TypeUtils to be able to compute QSymm8 of the weight * Create PerAxisIterator for per-axis quantization * Create QSymm8PerAxisDecoder * Create QSymm8PerAxisEncoder Signed-off-by: Keith Davis <keith.davis@arm.com> Change-Id: Ibcfe0288a197b7ee50b543bdbd77b7edb8a547c2
2019-09-10IVGCVSW-3824 Implement Float 16 Encoder and DecoderMatthew Jackson
* Implement Float 16 Encoder and Decoder * Add Stack Float 16 layer and create workload tests Signed-off-by: Matthew Jackson <matthew.jackson@arm.com> Change-Id: Ice4678226f4d22c06ebcc6db3052d42ce0c1bd67
2019-08-02IVGCVSW-3609 Fix decoding and encoding of INT32 tensorsAron Virginas-Tar
* Added Int32Decoder and Int32Encoder to decode INT32 tensors * Changed MakeDecoder to return ScaledInt32Decoder only if the scale is different from 0, i.e. for quantized bias tensors Signed-off-by: Aron Virginas-Tar <Aron.Virginas-Tar@arm.com> Change-Id: I278061d445d1c549c7ace11f51aa172ce7c691ae
2019-07-02IVGCVSW-3307 Don't assume TensorInfo::Map() can be called before Execute()Matthew Bentham
Change-Id: I445c69d2e99d8c93622e739af61f721e61b0f90f Signed-off-by: Matthew Bentham <Matthew.Bentham@arm.com>
2019-05-27IVGCVSW-3134 Refactor FullyConnected workloads into single workloadFrancis Murtagh
* Refactor FullyConnected workloads into single workload. * Refactor FullyConnected ref implementation to use Encoders and Decoders to support all DataTypes. * Deleted RefFullyConnectedFloat32Workload and RefFullyConnected2dUint8Workload. Change-Id: Iad30fb0287ab7491e1297997e7d61f1d00785541 Signed-off-by: Francis Murtagh <francis.murtagh@arm.com>
2019-05-23IVGCVSW-3025: Refactor reference Convolution2d workloadMike Kelly
* Refactored RefConvolution2dWorkload to support all DataTypes through Encoders and Decoders. * Added Convolute function to ConvImpl that uses Encoders and Decoders to support all DataTypes. * Deleted RefConvolution2dFloat32Workload and RefConvolution2dUint8Workload. Signed-off-by: Mike Kelly <mike.kelly@arm.com> Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com> Change-Id: Ic5ef0f499d08b948fa65fdee54b5f681fd0b1c05
2019-04-29IVGCVSW-2982 Refactor reference Activation workloadNattapat Chaimanowong
Change-Id: Ia3b9a56787cc68822a3c1635de82e03ecc0aae27 Signed-off-by: Nattapat Chaimanowong <nattapat.chaimanowong@arm.com>
2019-04-10IVGCVSW-2946 RefElementwiseWorkload configures prior to first executeDerek Lamberti
+ Added PostAllocationConfigure() method to workload interface + Elementwise function now deduces types based on Functor - Replaced RefComparisonWorkload with RefElementwiseWorkload specialization + Fixed up unit tests and minor formatting Change-Id: I33d08797767bba01cf4efb2904920ce0f950a4fe Signed-off-by: Derek Lamberti <derek.lamberti@arm.com>
generated by cgit v1.2.1 (git 2.23.0) at 2024-07-07 03:34:49 +0000