Gitiles
Code Review Sign In
review.mlplatform.org / ml / ComputeLibrary / 52016545c1df6af650f9ea1a00a7c2870fa50ce9 / src
  1. 4fbcac6 COMPMID-2102: Fix CLWinogradConvolutionLayer for 7x7 by Gian Marco Iodice · 5 years ago
  2. 86cfffe COMPMID-2099: Enable dummy threads in CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel by Gian Marco Iodice · 5 years ago
  3. 642680a COMPMID-1983 : Optimize NEIm2Col with NHWC data layout by Vidhya Sudhan Loganathan · 5 years ago
  4. 3bfacb2 COMPMID-1318: Implementing Winograd 7x7 NHWC on OpenCL - Part III by giuros01 · 5 years ago
  5. dbfc2dc COMPMID-2069: Rework CL ML layers to run exclusively on CL. by Georgios Pinitas · 5 years ago
  6. 881c684 COMPMID-1318: Implementing Winograd 7x7 NHWC on OpenCL -- Part II by Michele Di Giorgio · 5 years ago
  7. cafec8f COMPMID-1024 Investigate concatenation for RNN/LSTM OpenCL by John Kesapides · 5 years ago
  8. 62251f7 COMPMID-2002: Implement CLGEMMLowpMatrixMultiplyReshapedOnlyRHS - Transposed by Gian Marco Iodice · 5 years ago
  9. b4a44ff COMPMID-1995: Removed layout checks from Reduction ops by Pablo Tello · 5 years ago
  10. 4a8ec80 Optimize CL DeconvolutionLayer-Part II: Add CLDirectDeconvolution function to be used by CLDeconvolution. by giuros01 · 5 years ago
  11. cadb368 COMPMID-1995: Fixed graph fusion mutator for float types. by Georgios Pinitas · 5 years ago
  12. 0bc7849 COMPMID-1958: Implements 1D FFT in OpenCL. by Georgios Pinitas · 5 years ago
  13. 47d39dc COMPMID-1975: Update depthwise convolution. by Georgios Pinitas · 5 years ago
  14. 60f3911 COMPMID-2008: Add support for "reflect" padding mode in CLPad by Manuel Bottini · 5 years ago
  15. f955d51 COMPMID-1318: Implementing Winograd 7x7 NHWC on OpenCL -- Part I by Michele Di Giorgio · 5 years ago
  16. f52cd78 COMPMID-1995: Minor code fixes. by Georgios Pinitas · 5 years ago
  17. 8cf8c11 COMPMID-1944 Add support for "reflect" padding mode in NEPad by Usama Arif · 5 years ago
  18. ae0001e COMPMID-2077 Optimise CLL2NormalizeLayerKernel by Usama Arif · 5 years ago
  19. ec69975 COMPMID-2076: Add StackLayer to the graph API by Michele Di Giorgio · 5 years ago
  20. 2761c2f COMPMID-2075 Segfault in NELSTM by Michalis Spyrou · 5 years ago
  21. e214fcf COMPMID-2056: Remove data layout check in NECopyLayerKernel and CLCopyLayerKernel by Gian Marco Iodice · 5 years ago
  22. 6a14adb COMPMID-1932: Implement CLHeightConcatenateLayer. by Pablo Tello · 5 years ago
  23. 0ae5de9 COMPMID-1995: Prepare Graph to support different input/output quantization info by Isabella Gottardi · 5 years ago
  24. b0c5037 COMPMID-2043: Add support for "dummy threads" in CLGEMMReshaped by Gian Marco Iodice · 5 years ago
  25. 02acf01 COMPMID-2065: Fix validation issue in NEGEMMLowpMMCore by Georgios Pinitas · 5 years ago
  26. 1461383 COMPMID-1995: Update RSH GEMM assembly kernels. by Georgios Pinitas · 5 years ago
  27. 3dd5b68 COMPMID-1933: Implement NEHeightConcatenateLayer. by Pablo Tello · 5 years ago
  28. 2f7c149 Optimize CLDeconvolution-PART I: bug fixing CLGEMMLowpMatrixMultiplyCore by giuros01 · 5 years ago
  29. 6fa2638 COMPMID-1995: Fix compiler issues for invalid min,max arguments. by Georgios Pinitas · 5 years ago
  30. f7a3bf2 COMPMID-1995: Fix NEActivation Logistic by Isabella Gottardi · 5 years ago
  31. 2d7e683 COMPMID-1694: Fuse offset contribution with the output stage when we use NEGEMMLowpMatrixMultiplyCore by George Wort · 5 years ago
  32. 749021a COMPMID-1995: Revert fusing convolution to batch norm due to performance regressions by giuros01 · 5 years ago
  33. e03802e COMPMID-1936: Add support for QASYMM8 in CLQuantizeLayer. by Usama Arif · 5 years ago
  34. 917959c COMPMID-1281 Investigate concatenation for RNN/LSTM NEON by John Kesapides · 5 years ago
  35. 894066d COMPMID-1844: Implement CLCrop by George Wort · 5 years ago
  36. adfb273 COMPMID-1935 Add support for QASYMM8 in NEQuantizeLayer by John Kesapides · 5 years ago
  37. 381fcf2 COMPMID-1995: Allow calculations being performed in pixels borders in conv. by Georgios Pinitas · 5 years ago
  38. a42f55f COMPMID-1995: Allow weights and bias to be passed as SubStream in FullyConnectedLayer by Michele Di Giorgio · 5 years ago
  39. acce504 COMPMID-1740: Fuse batch normalization with Convolution Layer at graph level by giuros01 · 5 years ago
  40. ba5e096 COMPMID-1964: Implement CLGEMMMatrixMultiplyReshapedOnlyRHS - Not transposed by Gian Marco Iodice · 5 years ago
  41. efc4a2d COMPMID-1995: Fix call_all_input_node_accessors by Michele Di Giorgio · 5 years ago
  42. da0e513 COMPMID-1882: Improve memory coalescence when reshaping the weights for CLDepthwiseConvolution by Pablo Tello · 5 years ago
  43. 4710436 COMPMID-1882: Improve memory coalescence when reshaping the weights for CLDepthwiseConvolution by Pablo Tello · 5 years ago
  44. adc5395 COMPMID-2000: Implement CLGEMMMatrixMultiplyReshapedOnlyRHS - Transposed by Gian Marco Iodice · 5 years ago
  45. f00d332 COMPMID-2034 Compilation error in bare metal build by Manuel Bottini · 5 years ago
  46. 574775c COMPMID-1937: Adds support for DequantizationLayer for NEON/CL. by Georgios Pinitas · 5 years ago
  47. 79fa9a2 COMPMID-2009: Add support for QASYMM8 in NEPixelWiseMultiplicationKernel by Manuel Bottini · 5 years ago
  48. 05398a9 COMPMID-1843: Implement NECrop by George Wort · 5 years ago
  49. 6a2b6e8 COMPMID-2010: Add support for QASYMM8 in NEArithmeticSubtractionKernel by Manuel Bottini · 5 years ago
  50. fc1da13 COMPMID-1946: Implement NEBatchToSpace by giuros01 · 5 years ago
  51. fb68ca1 COMPMID-1490 Create graph_convolution.cpp by John Kesapides · 5 years ago
  52. ba36825 COMPMID-1947: Implement NESpaceToBatch by giuros01 · 5 years ago
  53. 9197c96 COMPMID-1995: Fix output quantization CLDeptwiseConv3x3 when activation is fused by Isabella Gottardi · 5 years ago
  54. 7097e3c COMPMID-1601: Graph examples segfaulting on some platform when gles=1 by Georgios Pinitas · 5 years ago
  55. 8f5802f COMPMID-1710: Add FP16 support checks. by Georgios Pinitas · 5 years ago
  56. 804aafb COMPMID-1710: Fix NEReductionOperation border values for QASYMM8. by Georgios Pinitas · 5 years ago
  57. 2481d46 COMPMID-1710: Quantization information missmatching issues. by Georgios Pinitas · 5 years ago
  58. d57891a COMPMID-2007: Compilation failures with ndk16b. by Georgios Pinitas · 5 years ago
  59. f85433b COMPMID-2006: NEON GEMMLowp assertion failure. by Georgios Pinitas · 5 years ago
  60. e46a7be COMPMID-2005: Tensors have different quantization information. by Georgios Pinitas · 5 years ago
  61. 5209be5 COMPMID-1999: Add support for GenerateProposals operator in CL by Manuel Bottini · 5 years ago
  62. 453ef52 COMPMID-1710 fix CL context creation by John Kesapides · 5 years ago
  63. 0a1090a COMPMID-1710: Add check on different input/output quantization info were not supported by Isabella Gottardi · 5 years ago
  64. e5bf4c5 COMPMID-1710: Fix CL logistic activation for QASYMM8 by Michele Di Giorgio · 5 years ago
  65. e96e4f0 COMPMID-1766: Implemented CPP Non Max Suppression by Pablo Tello · 6 years ago
  66. bb365de Revert "COMPMID-1329: Add support for GenerateProposals operator in CL" by Giuseppe Rossini · 5 years ago
  67. 5576315 Revert "COMPMID-1766: Implemented CPP Non Max Suppression" by Pablo Marquez · 5 years ago
  68. 6eb90d7 COMPMID-1996: Update documenation for doxgen 1.8.15 by Manuel Bottini · 5 years ago
  69. eb6c88a COMPMID-1918: Different qinfos support in CLConcatLayer by Pablo Tello · 5 years ago
  70. ca65af3 COMPMID-1919: Opencl - Tuning with Instrumentation does not work by Vidhya Sudhan Loganathan · 5 years ago
  71. 581c898 COMPMID-1934: Doxygen online documentation errors on Firefox by Manuel Bottini · 5 years ago
  72. 6844b88 COMPMID-1710 Fix NEON LSTM invalid allocation by Michalis Spyrou · 5 years ago
  73. 8e150a1 COMPMID-1859: Avoid unnecessary padding in CLWidthConcatenate[2,4]TensorsKernel by Michele Di Giorgio · 6 years ago
  74. 54e98d9 COMPMID-1918: Different qinfos support in NEConcatLayer. by Pablo Tello · 5 years ago
  75. 03bb550 COMPMID-1916: Add support for QASYMM8 LOGISTIC activation in NEActivationLayer by Isabella Gottardi · 5 years ago
  76. 4f85982 COMPMID-1923: Input is ignored if peephole optimization is enabled in LSTM by Georgios Pinitas · 5 years ago
  77. ebc3a90 COMPMID-1706: Fuse the bias addition within CLGEMM by Michele Di Giorgio · 6 years ago
  78. 4632e5e COMPMID-1920: Failure exit order destruction when CLTuner enabled. by Georgios Pinitas · 5 years ago
  79. a69a88b COMPMID-1915: Deconvolution doesn't work when inner_dimension_top != 0 or inner_dimension_right != 0 by giuros01 · 5 years ago
  80. 99089ce COMPMID-1924: Validate quantization info in pooling layer. by Georgios Pinitas · 5 years ago
  81. 678d83a COMPMID-1838: Add 4D softmax support for NEON and achieve parity with CL by Manuel Bottini · 5 years ago
  82. a74923c COMPMID-1910: Improve CLTuner reducing the number of LWS to test by Gian Marco Iodice · 5 years ago
  83. 3dbfd23 COMPMID-1710: Introduce GEMM strategy name in GEMMAssemblyWrapper. by Georgios Pinitas · 5 years ago
  84. 1509e4b COMPMID-1823: Increase scheduling granularity in NEGemmInterleaved. by Georgios Pinitas · 5 years ago
  85. a41c54b COMPMID-1710: Enable fp16 fallback for AArch32 by Georgios Pinitas · 5 years ago
  86. 6a91440 COMPMID-1710: Allow NHWC datalayout in SAME pad calculation by Isabella Gottardi · 5 years ago
  87. b7c308a COMPMID-1887: Fixed segfault in validation tests. by Pablo Tello · 5 years ago
  88. f8b6520 COMPMID-1911 : NEON QASYMM8 add should be rounding to nearest, tie to even by Vidhya Sudhan Loganathan · 5 years ago
  89. aa3fc13 COMPMID-1710: Re-enable winograd on A53. by Georgios Pinitas · 5 years ago
  90. 6d10996 COMPMID-1691: Optimize CLDepthwiseConvolutionKernel (QASYMM8/NHWC) for 3x3 kernels (stride=1 and stride=2) by giuros01 · 5 years ago
  91. edc21e4 COMPMID-1912: Print functions should be enabled in assert mode by giuros01 · 6 years ago
  92. 5d62c01 COMPMID-1710: Fix NEActivationLayer map for QASYMM8 by Isabella Gottardi · 5 years ago
  93. 83bbf77 COMPMID-1710: Define missed specializations for comparisons by Georgios Pinitas · 5 years ago
  94. db63b9c COMPMID-1698: Implementing CLGEMMLowpMatrixMultiplyReshapedKernel by Gian Marco Iodice · 5 years ago
  95. 17a01a3 COMPMID-1866: Revisit padding and window on CLDepthwiseConvolutionNHWC by Michele Di Giorgio · 6 years ago
  96. 1d48065 COMPMID-1867: Add u8 and s8 hybrid assembly kernels. by Georgios Pinitas · 5 years ago
  97. 20b527a COMPMID-1900: Nightly issue with GEMMReshapeLHSMatrix by Gian Marco Iodice · 5 years ago
  98. dc9b528 COMPMID-1795: Add support for NHWC in CLGenerateProposalsLayer by George Wort · 5 years ago
  99. 1d4f385 COMPMID-1760: NEON: Implement Prod by Manuel Bottini · 5 years ago
  100. 734151d COMPMID-1848: Account alignment in Offset-based pool allocations by Georgios Pinitas · 5 years ago