Age | Commit message (Collapse) | Author |
|
Change-Id: I3f3e247728fd6dafca066e41835f0ef9442d9b7a
Signed-off-by: giuros01 <giuseppe.rossini@arm.com>
Reviewed-on: https://review.mlplatform.org/c/1379
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
|
|
NEGEMMLowpMatrixMultiplyCore
Change-Id: Ic1a681e4cc03e1eba3bf8485d9cdb17b3e926047
Signed-off-by: giuros01 <giuseppe.rossini@arm.com>
Reviewed-on: https://review.mlplatform.org/c/561
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
|
|
Change-Id: Iba0739450c144dde141de1f2a3864cecc93c8266
Reviewed-on: https://review.mlplatform.org/527
Reviewed-by: Isabella Gottardi <isabella.gottardi@arm.com>
Tested-by: Arm Jenkins <bsgcomp@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>
|
|
Removed fixed point related code.
Change-Id: I487acf138dace3b0450e0d72ca7071eaec254566
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/137678
Tested-by: Jenkins <bsgcomp@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
|
|
Remove QS8 and QS16 validation and benchmark tests
Change-Id: I566f1474c1fafcb3903115ec2d3a003d73e4c93b
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/133762
Tested-by: Jenkins <bsgcomp@arm.com>
Reviewed-by: Pablo Tello <pablo.tello@arm.com>
|
|
Change-Id: I775eecbc39da583aae2eb4e033c5930dfc402899
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/125684
Tested-by: Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
|
|
Change-Id: I1e2a1a77097d8017c274af3f97eba6964f80f5fa
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/122592
Tested-by: Jenkins <bsgcomp@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
|
|
Used a left-over for loop in:
- NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel
- NEGEMMLowpQuantizeDownInt32ToUint8ScaleKernel
in order to remove the padding requirements for AndroidNN
Change-Id: I8ef529fc3d1adecf15fbe42002d99bc0030f131f
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/112266
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
Tested-by: BSG Visual Compute Jenkins server to access repositories on http://mpd-gerrit.cambridge.arm.com <bsgcomp@arm.com>
|
|
Change-Id: Ib57d4f7177cc6179302bda7ad870acb8bd3825f5
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/112115
Tested-by: BSG Visual Compute Jenkins server to access repositories on http://mpd-gerrit.cambridge.arm.com <bsgcomp@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
|
|
DoD:
- Implement NEON kernel for quantizing down the gemmlowp result. The
result should be scaled by a fixedpoint number
- Implement OpenCL kernel for quantizing down the gemmlowp result. The
result should be scaled by a fixedpoint number
- Add test for validating the result
Required for:
- Integration of GEMMLowp in Android NN
- Convolution quantized
- Fully connected quantized
Change-Id: Ia963d25d695471e963961fb49a5600e78374ac4f
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/110981
Tested-by: BSG Visual Compute Jenkins server to access repositories on http://mpd-gerrit.cambridge.arm.com <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
|
|
Reworked the interface of GemmLowp in order to make easy the integration
in Android NN
- Added support for different output stage
- Added validation for both matrix multiplication and output stage
- Added bounded relu support in the output stage
- Added in32_t bias support
- Added optimized path for vector by matrix case
This rework is required for:
- Convolution quantized
- Fully connected quantized
Change-Id: I512283d406099cf8c614dd89d0a97ed411143afc
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/110625
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Tested-by: BSG Visual Compute Jenkins server to access repositories on http://mpd-gerrit.cambridge.arm.com <bsgcomp@arm.com>
|
|
Integrate generic gemmlowp assembly version for u8.
Change-Id: I17ed4494c25a132b2bac581febe1544e49b4f352
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/110114
Tested-by: BSG Visual Compute Jenkins server to access repositories on http://mpd-gerrit.cambridge.arm.com <bsgcomp@arm.com>
Reviewed-by: Pablo Tello <pablo.tello@arm.com>
|
|
Change-Id: I8a470cc1351593ad8eeaf4ec92e04865e83d4f3c
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/96147
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
|
|
- Added min and max arguments for QuantizeDownInt32ToUint8Scale in order
to apply bounded relu
- Added support for int32_t biases
- Extended tests
Change-Id: I015dae17faa7284766b5435ca33bcf593c1b2b69
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/96512
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
|
|
Added support for S8 input in NEGEMMLowp Matrix Multiply Kernel.
Added a new function to run assembly kernels such that A*B=C (no offsets involved)
Added new tests for the assembly gemmlowp kernels (no offsets)
Integrated the assembly kernel for the A57
Change-Id: Ib3e39c1f3f7f1baa0d39be69485f61cd18e3c9b3
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/95864
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
|
|
The new interface makes NEGEMMLowp able to work with ASYMM8 data types.
Implemented 2 new functions:
- NEGEMMLowpMatrixMultiplyCore
- NEGEMMLowpOutputStage
These functions should make the integration in android NN doable
For more information about GEMMLowp:
https://github.com/google/gemmlowp/blob/master/doc/low-precision.md
Change-Id: Ie2c775f45234f68ca53dba644b3a912b997fd890
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/95504
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
Reviewed-by: Pablo Tello <pablo.tello@arm.com>
|
|
Change-Id: Id8f919e486a132fc58346c9f84fccbeeb83d19b3
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/94233
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
|
|
Change-Id: Ibbeff5d37249b6e8fc34ad496035a1511c9da5a3
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/94072
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
Reviewed-by: Pablo Tello <pablo.tello@arm.com>
|
|
Change-Id: Id02a0b3bf5af65dd940b46b2f6634b6a479cf388
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/92275
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
|
|
merged
Change-Id: Ie7bf46a157358ed64b4397002b10f583e7740c69
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/92367
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
|
|
Change-Id: I15496b16ffd636f5bff76572e750df7e15c80830
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/90532
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
|
|
Change-Id: I7bbe0f19a7255a37a6b0ce8c4d0bb57cf21c4bfb
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/89677
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
|