Age | Commit message (Collapse) | Author |
|
* ClSoftmaxKernel and ClSoftmax are created.
* ClSoftmaxKernel is now state-less and ClSoftmax handles
the internal tensors required for computation.
* add_const_tensor() is added to TensorPack not only
to have symmetric interface but also to benefit from
implicit conversion.
Implements: COMPMID-3998
Change-Id: I4f823121777be24260fd12b2cd71a6ff718c4eed
Signed-off-by: Sang-Hoon Park <sang-hoon.park@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5087
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
Resolves COMPMID-4151
Change-Id: I46f541efe8c4087f27794d2e158b6c1547d459ba
Signed-off-by: Michalis Spyrou <michalis.spyrou@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5160
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
|
|
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Change-Id: I4cc002da82da2219f3909a4e34463946cde4cf65
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5155
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
|
|
Change-Id: I5802c470683647b7426b3b6e7d17280cabc32163
Signed-off-by: Giorgio Arena <giorgio.arena@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5100
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michalis Spyrou <michalis.spyrou@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
- Revert changes in strides > num_dimensions. Set them to 0
- Fix offset calculcation in depthwise 3x3 quantized using select and stride_y for max offset
Resolve COMPMID-4254
Change-Id: Ia99b9637f18b99b1fa3d4b7b4892046027d3e7e5
Signed-off-by: Giorgio Arena <giorgio.arena@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5040
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
Fix errors when computing tensors with one element only
- Replace Tensor3D with raw pointers so to get rid of offset to first element for NCHW layout
- Add stronger out of bound constraints for NHWC layout
- Set the border size to the input's padding for NHWC
- Fill the strides == 0 with the largest stride, so to avoid accessing empty strides and multiplying by 0
Resolve COMPMID-4088
Change-Id: I751a4e6d7094b3c42306ff7f53af848fd35f19ac
Signed-off-by: Giorgio Arena <giorgio.arena@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5024
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Manuel Bottini <manuel.bottini@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
Port following functions:
- CLCopy
- CLFill
- CLPermute
- CLReshapeLayer
- CLCropResize
Resolves: COMPMID-4002
Signed-off-by: Sheri Zhang <sheri.zhang@arm.com>
Change-Id: I8392aa515aaeb5b44dab6122be6a795d08376d5f
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5003
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
|
|
- The ARM DOT macro was using wrong variables for performing the dot
product
- K0 could be a non power of 2 values when IFM was not a multiple of 16
- Refactor the test for direct convolution NHWC
Resolves COMPMID-4135, COMPMID-4155
Change-Id: I3a2dc89ef613ae20245cfc28e76ea36c55eaf81d
Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4962
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: TeresaARM <teresa.charlinreyes@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
|
|
* In CLDirectConvolution check for non-bias separately
Resolves: COMPMID-4214
Change-Id: I83c0688e9b48d059665bbc6e1f0f050a516132d6
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4980
Reviewed-by: Giorgio Arena <giorgio.arena@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
|
|
Resolves COMPMID-4000
Change-Id: I64878f93c033b4928fdefbb964c37c67fdecfaab
Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4971
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Manuel Bottini <manuel.bottini@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
|
|
Resolves COMPMID-4208
Change-Id: I61ca670134a005462ad0528a5aff9507a90860e7
Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4942
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
Resolves COMPMID-4004
Change-Id: I1dfe8bc52c1ff394ea208ba98b51033c738746a4
Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4922
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
|
|
Resolves COMPMID-4006
Change-Id: Iddc32b0b250142aac9a4a7b9dc0eef462d196025
Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4913
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Sang-Hoon Park <sang-hoon.park@arm.com>
|
|
- Fix ambiguosity with select in OpenCL
- Define a new macro for signed integer data type of the same input data type's size. This is needed because some ops (e.g. logical operators) in OpenCL work in this way
Resolves: COMPMID-4116, COMPMID-4110
Signed-off-by: Giorgio Arena <giorgio.arena@arm.com>
Change-Id: I560eda63fce24abd03d061f78f2f2ca951053fd0
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4898
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Change-Id: Icbe4e6a7c6732a59bdda0136af44c4852452dfd1
Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4891
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
Resolves COMPMID-3995
Change-Id: I84172bed20924f1d9ae3b4d14d7b321e9494296e
Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4887
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
|
|
- Pass the quantized zero value to the opencl kernel
Fixes COMPMID-3908
Change-Id: I6454c2e49f5b150a99178f2d72e0afa0a2990b54
Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4884
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Giorgio Arena <giorgio.arena@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
- Refactor direct convolution for NHWC
- Remove old kernels for NHWC
- Change the heuristic in CLConvolutionLayer.cpp. The new direct
convolution implementation is faster than FFT
Resolves COMPMID-3908
Change-Id: Iee15ce7b04e21847b6eaae5c6d3c1b18180e7efc
Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4876
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
|
|
By handling more general NxM blocks (where M and N can be 1,2,4,8,16)
instead of only 4x4, 8x8, 16x16 and managing corner left values with
partial stores
Resolves: COMPMID-3923
Change-Id: I49b1a560c8325e00e061bd04edcf55034d04dcd8
Signed-off-by: Manuel Bottini <manuel.bottini@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4780
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Reviewed-by: Giorgio Arena <giorgio.arena@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
- Cast the destination pointer to (__global DATA_TYPE*) when VEC_SIZE == 1 in range.cl
Resolves: COMPMID-3906, COMPMID-4093
Signed-off-by: Giorgio Arena <giorgio.arena@arm.com>
Change-Id: Ic0a334d98785ea434ed81f89dbe34e7674991f82
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4792
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
|
|
Use of proper vector size with boundary checking loads and stores
Resolves: COMPMID-3922
Change-Id: Ib631d499603b860fcfdbe3da903b866a125359a8
Signed-off-by: Manuel Bottini <manuel.bottini@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4789
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
Add padding checks in configure
Resolves: COMPMID-3914
Change-Id: Ia5be67283402d8811ceb3007be3a666ab502d775
Signed-off-by: Manuel Bottini <manuel.bottini@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4787
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
Resolves: COMPMID-3912
Change-Id: I1f8bd3bfec263ebfd70bc96f9183ccdc3089db13
Signed-off-by: Giorgio Arena <giorgio.arena@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4754
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Manuel Bottini <manuel.bottini@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
|
|
* Add -C flag to instruct preprocessor not to strip comments. This is to
prevent marker comments like '// fall through' that suppresses certain
warnings from being removed.
* Fix unused variable warnings.
* Add M_PI definition that's missing from certain toolchain standard
libraries.
Resolves COMPMID-4054
Change-Id: I1d641db668685d4b678f3d0efed84bfe9e630b4b
Signed-off-by: SiCongLi <sicong.li@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4692
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
|
|
Resolves COMPMID-3918
Change-Id: I970b1eaf2ae6f2f5a8cfc318cd1a3dfd3ba36fdb
Signed-off-by: Manuel Bottini <manuel.bottini@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4668
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Giorgio Arena <giorgio.arena@arm.com>
|
|
NCHW data layout
Fix border size for CLWinogradInputTransformKernel with NCHW data layout by setting it to the input's paddings. Add new the new validation shapes to the WinogradInputTransform's dataset
Resolves COMPMID-4042
Change-Id: Id93ac86e75c94ea3f2f35edcedebafada928f34a
Signed-off-by: Giorgio Arena <giorgio.arena@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4694
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Manuel Bottini <manuel.bottini@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
|
|
Resolves COMPMID-3905
Updates following kernels::
- CLDeconvolutionLayerUpsampleKernel
- CLDeconvolutionReshapeOutputKernel
- CLInstanceNormalizationLayerKernel
- CLMaxUnpoolingLayerKernel
- CLPermuteKernel
- CLQLSTMLayerNormalizationKernel
- CLReorgLayerKernel
- CLReverseKernel
- CLSpaceToBatchLayerKernel
- CLSpaceToDepthLayerKernel
- CLGenerateProposalsLayerKernel
- CLFFTDigitReverseKernel
- CLFFTRadixStageKernel
- CLFFTScaleKernel
- CLFillBorderKernel
- CLGatherKernel
- CLStridedSliceKernel
- CLBoundingBoxTransformKernel
Change-Id: I067ec670ff9cceadb1dfbf60dabef311a567d99a
Signed-off-by: Manuel Bottini <manuel.bottini@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4713
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Giorgio Arena <giorgio.arena@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
|
|
Change-Id: I07222a9eb03c785bb63414f581152267b133e9fc
Signed-off-by: Manuel Bottini <manuel.bottini@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4699
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
|
|
Resolves: COMPMID-4051
Change-Id: I0c0bf97212dd281c19d5081e6247e7dc0c23cd6b
Signed-off-by: Giorgio Arena <giorgio.arena@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4687
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
Upsample functions and kernels can be replaced with the Scale as they
provide same functionality
Partially resolves: COMPMID-3996
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Change-Id: Ic2f9ba352c183aa87d69d551d5c172d0f22119e8
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4679
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
|
|
- Fixed a bug that corrected the number of dimensions of a TensorShape for added trailing 1s
- Avoided adding offset_first_element for the Depthwise 3x3 NCHW OpenCL kernels, since it wouldn't align with the window which is based on the output
- Adjusted padding requirements along the x for Depthwise 3x3 NCHW. The kernel should always add 2 * dilation_(x/y) to the num_elems_read_x/y
- Adjusted the kernel's border_size given to the border handler at function level
- Added the dataset that previously made the tests fail
Resolves: COMPMID-4041
Change-Id: Ifab7d38b263f12173fcc96a5f0bd3375756c3c53
Signed-off-by: Giorgio Arena <giorgio.arena@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4673
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: SiCong Li <sicong.li@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
|
|
Adding BitwiseOperation enum class
Generalizing CL Bitwise kernels with a single CLBitwiseKernel
Removing CL padding from CLBitwiseKernel
Change-Id: I79cd79c1e425b6da7d52308a420edf8cfb7a5a36
Signed-off-by: Manuel Bottini <manuel.bottini@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4646
Reviewed-by: Giorgio Arena <giorgio.arena@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
YOLO layer is too specialized and specific to a single model type. Can
be decomposed using split, activation and concatenate layers
Partially Resolves: COMPMID-3996
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Change-Id: I3cde88f8d4cc7d8c70ce1bb3b32b00f8d09bdca2
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4678
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
|
|
Flatten layer is lowered into a Reshape layer.
Remove (CL/NE)FlatternLayerKernel.
Partially Resolves: COMPMID-3996
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Change-Id: Id9e2ddfe2e2dd793541badff3490c05e4c908f88
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4660
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
|
|
Remove out-of-date and unmaintained LocallyConnectedLayer for both NEON
and OpenCL.
Resolves: COMPMID-3924
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Change-Id: Ia61398ed8cfa3876f41c1b342c4a80d1cca0ca83
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4634
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
Partially Resolves: COMPMID-3924
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Change-Id: Ibc47bd5bf5203dbad8d0755608918fcb384053c3
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4633
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
- LEAKY RELU activation is supported for QASYMM8 data type
- vquantize on NEON side has been modified to match with
other backends (OpenCL and reference)
Change-Id: I194631225c8d4f3cc96027d64812ec2be2b4328a
Signed-off-by: Sang-Hoon Park <sang-hoon.park@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4593
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Reviewed-by: Manuel Bottini <manuel.bottini@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
The mismatches are caused by out of bound memory access on weight tensor
due to lack of padding in the channel (first in NHWC) dimension.
Signed-off-by: SiCong Li <sicong.li@arm.com>
Change-Id: I5a73f190f8e131c67ed7769f6f716db9d79dc674
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4628
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Manuel Bottini <manuel.bottini@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
Change-Id: Id2cc77508b0f2fa36a298059476b01704cfbdcaf
Signed-off-by: Manuel Bottini <manuel.bottini@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4580
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
|
|
Change-Id: I04cd23e9abcb1828e54cd59fee3bfa95a6dea3fb
Signed-off-by: Sang-Hoon Park <sang-hoon.park@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4461
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Reviewed-by: Manuel Bottini <manuel.bottini@arm.com>
|
|
Change-Id: I612aeed6affa17624fb9044964dd59c41a5c9888
Signed-off-by: Sang-Hoon Park <sang-hoon.park@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4448
Reviewed-by: Pablo Marquez Tello <pablo.tello@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
* Add missing padding immutability asserts in all relevant CL kernels
* Remove unnecessary zero padding validation tests.
Change-Id: If93f9ccbc988e0286f5e7b135f812141476d5da0
Signed-off-by: SiCong Li <sicong.li@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4446
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
|
|
when fusing Bounded ReLU in Android R GpuAcc
Change-Id: I6cfee002846d0c84de7e0a5f141dfc4807b93b33
Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4421
Reviewed-by: TeresaARM <teresa.charlinreyes@arm.com>
Reviewed-by: Pablo Marquez Tello <pablo.tello@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
Change-Id: Iba02375df47d227feca07cc0215e3389e7c55ade
Signed-off-by: Manuel Bottini <manuel.bottini@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4401
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
- Renamed SELECT_DATA_TYPE to SELECT_VEC_DATA_TYPE to reflect its usage with vectors. SELECT_DATA_TYPE(dt) will now return the primitive data type
- Changed the interface of VEC_OFFS and V_OFFS in order to receive the primitive data type as a parameter rather than its vector form
- Performed a general cleanup of the kernels, such as creating macro for sum and max reduces, remove reduntant macros, defines, variables, calculations, etc...
- Using VEC_SIZE and VEC_SIZE_LEFTOVER in every kernel in order to allow computation for smaller shapes without adding paddings
- Removed the actual padding from the kernel and adjusting its calculations accordingly. Added asserts for padding removal checks. Removed invalid Validate tests.
Change-Id: If5ccbd5d34e255d38c7f6bfe8740e2b80b28e264
Signed-off-by: Giorgio Arena <giorgio.arena@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4277
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: SiCong Li <sicong.li@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
|
|
* Fix CLSpaceToBatchLayerKernel and NESpaceToBatchLayerKernel validation
errors by using the correctly calculated output tensor shape
Signed-off-by: SiCong Li <sicong.li@arm.com>
Change-Id: I21d61f870e6a23a2e38dcb95c939b0bf08082b6f
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4347
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Reviewed-by: TeresaARM <teresa.charlinreyes@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
Change-Id: I56137938c9ebe1a5aeeaa750b39fcfc6818016f1
Signed-off-by: SiCong Li <sicong.li@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4332
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
Change-Id: I10d27db788e5086adae1841e3e2441cd9b76ef84
Signed-off-by: Sang-Hoon Park <sang-hoon.park@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4310
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
|
|
* Remove default definition for STORE_BLOCK_BOUNDARY_AWARE to avoid elusive bugs
* Clean up gemm_mm_interleaved* and gemm_mm_floating_point* kernels
* Relocate to gemm_v1.cl to avoid clashing with new kernels
* Rename compile time arguments to conform with the established
terminology(MNKB), and to facilitate the use of STORE_BLOCK_BOUNDARY_AWARE
Change-Id: Ia85c746b2536cad87257a79685b459b5d2f9a1be
Signed-off-by: SiCong Li <sicong.li@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4329
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|
|
- Update heuristic for GEMM reshaped RHS only
- Fix left-over block size in CLGEMMMatrixMultiplyReshapedOlyRHSKernel
Change-Id: I34c738821ed2e4a537da4a15058eec164cb6b61f
Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4305
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
|