Age | Commit message (Collapse) | Author |
|
kernels
Change-Id: I98183f95814442b6f3dbb67a1bdae99df05b9b01
|
|
- Fixing a bug for which we did not scale the boxes before transforming them
- Adding the correct_transform_coords option to BoundingBoxTransformInfo
Change-Id: I40281254bcf87e7c8583c119e99562414fe59822
|
|
BoxWithNMSLimitKernel
COMPMID-1792: Accuracy issue in CLGenerateProposals
This patch does the following:
- Some fixes for GenerateProposals function and tests
- Adapting BoxWithNMSLimitKernel to only accept U32 tensors as keeps_size
- Update 3rdparty
- Adds a small tolerance for a GenerateProposals test
Change-Id: Ia8ec1cdfe941fe05003645e86deb9ea6a6044d74
|
|
Introduced F32 accumulation for F16 winograd gemm and output transform
WinogradConvolution will be available for F16 only if fast math flag is enabled
Change-Id: I215593c205236a0f9669218437bb40b184ec6a4f
|
|
Increase tolerance for FP16
Change-Id: I88f95da5471bbceb7449f453e2e33cf0bc4da23e
|
|
Change-Id: I99e1c3939cfea4b9cb0ddfa313706f31b213ca89
|
|
Change-Id: I69e995973597ba3927d29e4f6ed5438560e53d77
|
|
In case of CIFG optimisation scratch buffer should have a size of
[batch_size, num_units * 3] else [batch_size, num_units * 4].
Change-Id: I43e46f7b52e791472f1196f36e9142240ba76c5c
|
|
Added test cases to exercise the code path where the reshaping of B is performed on the fly.
Change-Id: Ifa4348e1054dc0019be3927f482adf64b18fd554
|
|
Change-Id: Ib0798cc17496b7817f5b5769b25d98913a33a69d
|
|
Change-Id: I5bf5d751ec7c02d96c26a769f49d03ea23a248b7
|
|
Change-Id: Ibab049f09413258c99335b7da6b151530a1bd136
|
|
and 8 tensors (Part 1)
Creating special cases for concatening 2 and 4 tensors.
Change-Id: I6a739a494ae45011acb65369e353f9ef96970b90
|
|
inside the namespace
Change-Id: I477f52a9adf06ba3730f94d411399977fce0f98a
|
|
the test.
This is needed in order to calculate the offset between OpenCL timestamps and Wall Clock timestamps as they're using different clocks
Change-Id: I874b2a475bf98fd664a1e3e15045c80f0181af47
|
|
duration
Change-Id: Iafc1d6cd8003de64a3439ad807f4002036c73a73
|
|
The issue was related to CLIm2Col when the number of input channels was less than
the number of elements processed by each thread.
The bug has been fixed in the validate_and_configure_window() function setting the correct number of elements accessed
in the output tensor.
Also fixed an issue GEMM3D when we have a single output channel
Change-Id: I094292d0c7662599c4a4c3916ec5f5821df5faef
|
|
Change-Id: I86679adff556b6ffc9929b35cbf1b59b3958bdb1
|
|
Change-Id: I6d5f91579850906e1eb973ff6c5612195255e631
|
|
Change-Id: I807ef84dbf893bd401dcac5c0fa3a4ee49aabc66
|
|
ArmNN reported an issue with padding in CLLSTMLayer. This was due to the fact
that some tensors were allocated before they were passed to some configure
functions which attempted to change the padding requirement on already allocated
memory.
Also, increase tolerance on number of mismatches for CLBBoxTransform FP16.
Change-Id: Iad75b012be895693d0e553f3ab85f1ca7144e882
|
|
The tab characters were corrupting the output JSON file of arm_compute_validation
Change-Id: I8792fd0e02393aef60341552b428111e969a3927
|
|
Reduce the amount of precommit tests run in DirectConvolution,
Deconvolution and Pooling. Proper investigation scheduled for later.
Change-Id: Idc2510cf6877e7a605cead84f384852b609e3216
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/156466
Tested-by: bsgcomp <bsgcomp@arm.com>
Reviewed-by: Vidhya Sudhan Loganathan <vidhyasudhan.loganathan@arm.com>
|
|
Note: Only ComputeLibrary files get copied over (Stub CL / GLES drivers don't, nor are the 3rdparty includes)
utils/ files are not copied either (They're not part of the core library)
Change-Id: I55e01c0ba4a5f7e649877fcdd11fdb0a51071b18
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/156339
Reviewed-by: Pablo Tello <pablo.tello@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Tested-by: bsgcomp <bsgcomp@arm.com>
|
|
Also added the test case reported by ArmNN.
Change-Id: I9fe9a1b4f74267a3346529f3a597b37486593c4a
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/155914
Tested-by: bsgcomp <bsgcomp@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
|
|
Change-Id: Iac6a95ba7f388e65b7f1c8865c3e9bf289b233ea
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/155490
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
Tested-by: bsgcomp <bsgcomp@arm.com>
|
|
Change-Id: I3c3e96a743614af4c2c2391780d5de2db6191b0f
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/155318
Tested-by: bsgcomp <bsgcomp@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
Reviewed-by: Pablo Tello <pablo.tello@arm.com>
|
|
OpenCL
COMPMID-1424 - Add dot product support for CLDepthwise QASYMM8 3x3 NHWC non-unit stride
With this patch we are able to improve the performance of MobileNet v1-qasymm8 by 37 %
Tried to use the dot product instruction in CLDepthwise QASYMM8 3x3 NHWC non-unit stride
but I have not seen any benefit (maybe because we have few arithemtic operation and we
do not have more load instructions). However Depthwise convolution has been improved by
30%
Change-Id: Id768a99c2e53a04276707e427af5d0ec93419ada
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/155082
Tested-by: bsgcomp <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
|
|
Clear CLContext in a more regular basis to make the driver release
memory back to the system.
Change-Id: I0df847766f57719433bbaeada45fe630e38c9541
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/155435
Reviewed-by: Giuseppe Rossini <giuseppe.rossini@arm.com>
Tested-by: bsgcomp <bsgcomp@arm.com>
|
|
Increases relative tolerance slightly as error was quite small.
Change-Id: I4789c5e3eeb4f2d3aaf2b4c76966474f045af4c1
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/155418
Reviewed-by: Michalis Spyrou <michalis.spyrou@arm.com>
Tested-by: bsgcomp <bsgcomp@arm.com>
|
|
COMPMID-1690: Add tests for NEPermute with PermutationVector dimension > 3
Change-Id: I4bfc6ff88cd46863c2e39975b5663c624db1a63d
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/155316
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Tested-by: bsgcomp <bsgcomp@arm.com>
|
|
Change-Id: I875ffe0ccec3aa4f53bfb68d82e2a7292ab83358
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/155348
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Tested-by: bsgcomp <bsgcomp@arm.com>
|
|
Set input range to [-1, 1] in order to avoid inf values
when calculating sqrt.
Change-Id: I18f1e427baa7830fdc587bedf27a92d78c72f49b
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/155397
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Tested-by: bsgcomp <bsgcomp@arm.com>
|
|
Change-Id: I7a32becd78fc231d11d50c6ff58892f4acb0ccda
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/155224
Tested-by: bsgcomp <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
|
|
Instead of changing the tolerances I increased the sizes of the input. In this way, for a single mismatch, as it was the case, we are below the 1% tolerance set.
Change-Id: I787261a1d1adb559c1687b7bd1e0317a72594130
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/155168
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Tested-by: bsgcomp <bsgcomp@arm.com>
|
|
Increase tolerance for fp32 and added absolute tolerance
Change-Id: Iff828457b514d6301ed2c8e04b66ce86867f72b6
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/155086
Tested-by: bsgcomp <bsgcomp@arm.com>
Reviewed-by: Isabella Gottardi <isabella.gottardi@arm.com>
|
|
NEWidthConcatenateLayerKernel works with 4D tensors too, hence the check has
been removed and tests have been added.
Change-Id: I73814cabe5fae975a44cc1a03b092c552497e57d
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/155070
Tested-by: bsgcomp <bsgcomp@arm.com>
Reviewed-by: Isabella Gottardi <isabella.gottardi@arm.com>
|
|
Change-Id: I7bd4a8ce81483ba56686b765ca3caabebe42882d
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/155000
Tested-by: bsgcomp <bsgcomp@arm.com>
Reviewed-by: Michele DiGiorgio <michele.digiorgio@arm.com>
|
|
vector
When one of the operands is a vector, the kernel does a broadcast addition and
the window is not collapsed. This represent an issue because it leads to a lot
of enqueues that increases the time taken by the OpenCL driver. This patch
allows to collapse the window when one of the two operands is a vector.
Furthermore, it adds LWS tuner to the kernel.
It also changes the number of elements processed per iteration to 8 to make
better usage of the cache.
Change-Id: I5f09ab0ddcffb3b7f9326a987c79a997b2d7fa8c
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/155003
Reviewed-by: Giuseppe Rossini <giuseppe.rossini@arm.com>
Tested-by: bsgcomp <bsgcomp@arm.com>
|
|
Change-Id: Ic8312a5b6790aa7cd4468d42f08d557ad40e9441
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/154570
Tested-by: bsgcomp <bsgcomp@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
|
|
Change-Id: I91865506166951b3bf7f06a0b2d4cde925cfefb6
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/153447
Tested-by: bsgcomp <bsgcomp@arm.com>
Reviewed-by: Pablo Tello <pablo.tello@arm.com>
|
|
Change-Id: Iae22554d5fe893fd22a000eab5bfd8275ea06eb3
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/154102
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Tested-by: bsgcomp <bsgcomp@arm.com>
|
|
Change-Id: I146936c9e98b343496a4b61cdbadf0eaa38e885a
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/154008
Reviewed-by: Michele DiGiorgio <michele.digiorgio@arm.com>
Reviewed-by: Giuseppe Rossini <giuseppe.rossini@arm.com>
Tested-by: bsgcomp <bsgcomp@arm.com>
|
|
Change-Id: Ibc0b1242804c2fdb183825406e3c78bd0d1d3564
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/154368
Reviewed-by: Pablo Tello <pablo.tello@arm.com>
Tested-by: bsgcomp <bsgcomp@arm.com>
|
|
Change-Id: Id974efad304c2513b8824a6561ad45ee60b9e7fb
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/153763
Reviewed-by: Giuseppe Rossini <giuseppe.rossini@arm.com>
Reviewed-by: Isabella Gottardi <isabella.gottardi@arm.com>
Tested-by: bsgcomp <bsgcomp@arm.com>
|
|
Change-Id: If496709958bf29589601eac62a268819736a4fd2
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/154173
Reviewed-by: Pablo Tello <pablo.tello@arm.com>
Tested-by: bsgcomp <bsgcomp@arm.com>
|
|
COMPMID-1651: Fix QASYMM8 CLDeconvolutionLayer
This patch also extends the range of values used for testing Convolution and
Deconvolution to cover quantized [-1.0f, 1.0f].
Change-Id: I8b280669db67bb3ec25bf5d411c8f5954f5b0dab
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/149869
Reviewed-by: Michalis Spyrou <michalis.spyrou@arm.com>
Tested-by: bsgcomp <bsgcomp@arm.com>
|
|
Change-Id: Id331199f569f52a37280a9ada5bf84694580b93c
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/152843
Tested-by: bsgcomp <bsgcomp@arm.com>
Reviewed-by: Michele DiGiorgio <michele.digiorgio@arm.com>
|
|
Change-Id: I05a1b871746a32ccc1c3ecec97b8266767c9d0a7
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/153715
Reviewed-by: Pablo Tello <pablo.tello@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Tested-by: bsgcomp <bsgcomp@arm.com>
|
|
Change-Id: I2c4dcedcd3b56e41174eebbbacd47be4e968d34d
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/152767
Tested-by: bsgcomp <bsgcomp@arm.com>
Reviewed-by: Vidhya Sudhan Loganathan <vidhyasudhan.loganathan@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
|