Gitiles
Code Review
Sign In
review.mlplatform.org
/
ml
/
ComputeLibrary
/
87e0e4d7b662e6bf030d95fd75e55ee3c66a3e80
« Previous
e413d25
COMPMID-1783: NEGEMMConvolutionLayer fails for QASYMM8 on 32bit
by Georgios Pinitas
· 6 years ago
e059f4f
COMPMID-1787: Change the heuristic selection in CLGEMMLowpMatrixMultiplyCore
by Gian Marco Iodice
· 6 years ago
47e6fed
COMPMID-1639: Add GenerateProposals to the graph API
by Michele Di Giorgio
· 6 years ago
80a2990
COMPMID-1784: Fixed mapx/mapy access windows
by Anthony Barbier
· 6 years ago
ee50f51
COMPMID-1784 Fixed the padding checks in AccessWindowRectangle::update_window_if_needed()
by Anthony Barbier
· 6 years ago
1056595
COMPMID-1737: Fix: forgot to copy the 'support' folder to the include folder in the install_dir
by Anthony Barbier
· 6 years ago
60f0a41
COMPMID-1676: Change CLROIAlign interface to accept ROIs as tensors
by Manuel Bottini
· 6 years ago
0cc37c3
COMPMID-1451: Fix the shape of scratch_buffer in case of CIFG
by Georgios Pinitas
· 6 years ago
0e37b5c
COMPMID-1708: Improve GEMM test coverage.
by Pablo Tello
· 6 years ago
cd96a26
COMPMID-1329: Add support for GenerateProposals operator in CL
by giuros01
· 6 years ago
1c32bf396
COMPMID-1451: Perform fusion before GroupConvolution unrolling
by Georgios Pinitas
· 6 years ago
6c7c38e
COMPMID-1462 SSD support: Create CL PriorBox
by Michalis Spyrou
· 6 years ago
d8e340d
COMPMID-1638: Add BBoxTransform to the graph API (3RDPARTY_UPDATE)
by Manuel Bottini
· 6 years ago
5538d34
COMPMID-1781 Add channel support in CLL2Normalization
by Michalis Spyrou
· 6 years ago
27400b9
COMPMID-1707: Create 3 special CLWidthConcatenate kernel to concatenate 2/4 and 8 tensors (Part 1)
by Michele Di Giorgio
· 6 years ago
bb081ca
COMPMID-1751: Remove output_3d_depth from NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint
by Georgios Pinitas
· 6 years ago
cc225be
COMPMID-1777 Fixed compilation error due to templates being instantiated inside the namespace
by Anthony Barbier
· 6 years ago
9c2ec7e
COMPMID-1451: Fix clang-tidy issues.
by Georgios Pinitas
· 6 years ago
2e8a2dc
COMPMID-1451: Set axis correctly in CLL2Normalize validate function
by Georgios Pinitas
· 6 years ago
e7f4a43
COMPMID-1777: Output 'now' measurements in OpenCLTimer at the beginning of the test.
by Anthony Barbier
· 6 years ago
0cf7798
COMPMID-1626: Fixed VGG 16/19 bad_alloc failure.
by Pablo Tello
· 6 years ago
283fc60
COMPMID-1451: Reduces accuracy issue in NEPoolingLayer for QASYMM8 NHWC
by Georgios Pinitas
· 6 years ago
742df6c
COMPMID-1451 - Removed unused OpenCL kernel from gemmlowp.cl
by Gian Marco Iodice
· 6 years ago
027ce5b
COMPMID-1451: Perform two steps in calculating sqrt in NEPoolingLayer
by Georgios Pinitas
· 6 years ago
1e92896
COMPMID-1451: Dispatch fp16 gemms through the function interface.
by Georgios Pinitas
· 6 years ago
54c92fd
COMPMID-1451: Remove unused code.
by Georgios Pinitas
· 6 years ago
cfcc41c
COMPMID-1675: Remove unused SVE files
by Georgios Pinitas
· 6 years ago
c73e2b8
COMPMID-1451: Fix fused activation in GEMMConvolutionLayer
by Georgios Pinitas
· 6 years ago
5e25b12
COMPMID-1451: Allow weights retention in CLDeconvolutionLayer
by Michele Di Giorgio
· 6 years ago
72f4ae5
COMPMID-1777: Add option to make instruments output timestamps instead of duration
by Anthony Barbier
· 6 years ago
d2048ce
COMPMID-1638 - Add BBoxTransform to the graph API
by Manuel Bottini
· 6 years ago
0c54a62
COMPMID-1451: Removed output_depth3d from CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFloat
by Gian Marco Iodice
· 6 years ago
3139f03
COMPMID-1736: Fixed out-of-bound write in CLIm2Col
by Gian Marco Iodice
· 6 years ago
421405b
COMPMID-1675: Add SVE support
by Georgios Pinitas
· 6 years ago
f1adf11
COMPMID-1579: Add support for ChannelShuffle operator in NEON
by Georgios Pinitas
· 6 years ago
83e3e75
COMPMID-1776: Revert QuantizeDownStage to use fixed-point
by Georgios Pinitas
· 6 years ago
a518f30
COMPMID-1773: Replaced Gemm wrapper for NEGEMM in NEWinograd
by Pablo Tello
· 6 years ago
bdbbbe8
COMPMID-1744: Add paper references in graph models.
by Georgios Pinitas
· 6 years ago
d636bc5
COMPMID-1451: Fixed zerobuff sizes and clobbers in interleave transforms.
by Georgios Pinitas
· 6 years ago
dd2619a
COMPMID-1451: Fix order of allocations in CLLSTMLayer
by Michele Di Giorgio
· 6 years ago
2709d61
COMPMID-1328 Add support for BoxWithNMSLimit operator in CPP
by Michalis Spyrou
· 6 years ago
3704464
COMPMID-1703: Collapse the 4th dimensions in CLDepthWiseConvolutionLayer3x3Kernel
by Georgios Pinitas
· 6 years ago
29421bd
COMPMID-1743: Removed tabs from compare_keypoints' output
by Anthony Barbier
· 6 years ago
b7cb536
Added README and LICENSE, updated gitmodules to point at internal Arm server
by Anthony Barbier
· 6 years ago
064add6
COMPMID-1451 Reduce precommit tests
by Michalis Spyrou
· 6 years ago
f8defca
COMPMID-1451 Fix CLL2normalize typos
by Michalis Spyrou
· 6 years ago
5784624
COMPMID-1739: Fix broadcast CLArithmeticAddition for QASYMM8
by Michele Di Giorgio
· 6 years ago
01bbd5f
COMPMID-1737: Add support for install_dir
by Anthony Barbier
· 6 years ago
645e837
COMPMID-1712 CLPoolingLayer wrong results in QASYMM8
by Michalis Spyrou
· 6 years ago
9da19e9
COMPMID-1605: API alignment for the MemoryManager with ARMNN
by Georgios Pinitas
· 6 years ago
ffb57a0
COMPMID-1699: Disable arithmetic operations in CLWinogradLayer when no batches available.
by Georgios Pinitas
· 6 years ago
9593bde
COMPMID-1674: (3RDPARTY_UPDATE) Add FullyConnectedLayer Dragonbench
by Georgios Pinitas
· 6 years ago
89d7173
COMPMID-1704: Collapse the 4th dimension in CLPoolingLayerKernel
by Georgios Pinitas
· 6 years ago
f718ce2
COMPMID-1701: small cleanup NEWinogradConvolutionLayer.
by Pablo Tello
· 6 years ago
4b90865
COMPMID-1413 - Improve the performance of GEMMLowp with 8 bit dot product on OpenCL
by Gian Marco Iodice
· 6 years ago
c55beee
COMPMID-1029: Collapse CLWinogradInputTransform/CLWinogradOutputTransform
by Georgios Pinitas
· 6 years ago
f5ec981
COMPMID-1695: (Nightly) bad_alloc in CLIm2Col
by Georgios Pinitas
· 6 years ago
009fede
COMPMID-1696: (Nighlty) CLDepthwiseConvolution FP16 mismatches
by Georgios Pinitas
· 6 years ago
4ad6598
COMPMID-1451: Fix CL/NEPermuteKernel PermuteVection check
by Isabella Gottardi
· 6 years ago
e8a55df
COMPMID-1685: Allow build_dir to be an absolute path
by Anthony Barbier
· 6 years ago
e55a013
COMPMID-1451: Fix validation issue in CLReduceMean
by Michalis Spyrou
· 6 years ago
d775cd7
COMPMID-1451: FP16 L2Normalization failures
by Michalis Spyrou
· 6 years ago
13404bf
COMPMID-711: add GEMMInterleave4X4 benchmarking tests.
by Pablo Tello
· 6 years ago
af7f740
COMPMID-1451: Fix inlines in cl helpers
by Georgios Pinitas
· 6 years ago
201c37c
COMPMID-1530 error: dereferencing type-punned pointer will break strict-aliasing rules
by Michalis Spyrou
· 6 years ago
517055b
COMPMID-1451: Remove invalid bias check in NEDeconvolutionLayer
by Georgios Pinitas
· 6 years ago
cccb3c7
COMPMID-1451: DATA_UPDATE Update data folder
by Michalis Spyrou
· 6 years ago
a6bc82f
COMPMID-1680: (Nighlty) CLBBoxTransform mismatches
by giuros01
· 6 years ago
0bebac5
COMPMID-1682 Nightly NEON Reduction operation mismatches
by Michalis Spyrou
· 6 years ago
d7cd361
COMPMID-1681: (Nightly) NEWidthConcatenateLayer fails
by Michele Di Giorgio
· 6 years ago
96f8461
COMPMID-1451 Properly remove dimensions in CLReduceMean
by Michalis Spyrou
· 6 years ago
a1422fb
COMPMID-1673: Collapse window in CLArithmeticAddition when one operand is a vector
by Michele Di Giorgio
· 6 years ago
33893c3
COMPMID-1641: Create Facebook's MaskRCNN v2 graph example (3RDPARTY_UPDATE)
by Georgios Pinitas
· 6 years ago
51e53a3
COMPMID-1451: Perform CLOutputStage using floats.
by Georgios Pinitas
· 6 years ago
60e9825
COMPMID-1451: Fuse activation in DepthwiseConvolution.
by Georgios Pinitas
· 6 years ago
c04a0e8
COMPMID-1327: Add support for BBoxTransform operator in CL
by giuros01
· 6 years ago
08346e9
COMPMID-1451:Fuse RELU,LU_BOUNDED_RELU with requantization in NEGEMMConvolutionLayer.
by Georgios Pinitas
· 6 years ago
8aaf93e
COMPMID-1632 Add CLL2NormalizationLayer for NHWC and FP32
by Michalis Spyrou
· 6 years ago
c936917
COMPMID-1523: Fuse BN node with convolution.
by Georgios Pinitas
· 6 years ago
e6dbde0
COMPMID-1667: Add 4D tensors support to CLWidthConcatenateLayerKernel
by Michele Di Giorgio
· 6 years ago
bcf8a96
COMPMID-1580 Implement ReduceMean in NEON
by Michalis Spyrou
· 6 years ago
089695f
COMPMID-1659: (Nightly) CLGEMMConvolutionLayer QASYMM8 TensorShape error
by Isabella Gottardi
· 6 years ago
20ff84d
COMPMID-1451 - Enabled NHWC in AlexNet, InceptionV3 and InceptionV4
by Gian Marco Iodice
· 6 years ago
7c75ba0
COMPMID-1661 Nightly - CLReductionOperation FP16 failures
by Michalis Spyrou
· 6 years ago
061dd36
COMPMID-1451: Fix allocation of weights in Deconvolution
by Michele Di Giorgio
· 6 years ago
ed5a492
COMPMID-1586: Add support for NHWC CLDeconvolutionLayer
by Michele Di Giorgio
· 6 years ago
7e9391b
COMPMID-1574 Implement ReduceMean in OpenCL
by Michalis Spyrou
· 6 years ago
555c3d6
COMPMID-1657 - Fix CPU info in the json file output
by Gian Marco Iodice
· 6 years ago
c4f582e
COMPMID-1451: Reverting changes for CLGEMM and CLGEMMLowp previuosly done (384496)
by Isabella Gottardi
· 6 years ago
21eaefe
COMPMID-1631: Fixed test framework memory corruption.
by Pablo Tello
· 6 years ago
7f32d01
COMPMID-1451: Fix NormalizationLayer accross width normalization.
by Georgios Pinitas
· 6 years ago
afbc5ff
COMPMID-1621 Deconvolution wrong output calculation
by Michalis Spyrou
· 6 years ago
67d94d2
COMPMID-1451: Fix compilation issues under gcc 8
by Georgios Pinitas
· 6 years ago
af7e600
COMPMID-1628: NEWinograd mismatches in the nightly.
by Pablo Tello
· 6 years ago
d3d97d2
COMPMID-1623: NEWinograd reduce the number of output tiles.
by Pablo Tello
· 6 years ago
709d27b
COMPMID-1622 : (Nightly) Seg fault on android
by Vidhya Sudhan Loganathan
· 6 years ago
ecae3a1
COMPMID-1451: Enable dot kernels in NEGEMMAssembly functions
by Georgios Pinitas
· 6 years ago
df3b5bb
COMPMID-1630: Fixed broken DOTPROD detection.
by Pablo Tello
· 6 years ago
e95c206
COMPMID-1622 : (Nightly) Seg fault on android
by Vidhya Sudhan Loganathan
· 6 years ago
f02e527
COMPMID-1607 - (Nightly) CLGEMMLowpMatrixMultiplyCore errors and mismatches
by Isabella Gottardi
· 6 years ago
Next »