Commits

Branch selector

 9a59fbd

User selector

All users

Datepicker

All time

Commit History

Commits on Jan 14, 2019

• FP16Benchmark: Allow fp32 comparison using cblas (pytorch#56)

  Show description for 9a59fbd
  

  WilliamTambellini

  authored and

  facebook-github-bot

  committed
  9a59fbd

  Copy full SHA for 9a59fbd

Commits on Jan 12, 2019

• 3x3x3 depthwise convolution with per channel quantization (#15775)

  Show description for 36309fc
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  36309fc

  Copy full SHA for 36309fc

Commits on Jan 11, 2019

• don't keep conv_param_p member as a const reference (pytorch#57)

  Show description for f6e4f99
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  f6e4f99

  Copy full SHA for f6e4f99

Commits on Jan 4, 2019

• missing copyright headers

  Show description for a7ac276
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  a7ac276

  Copy full SHA for a7ac276

Commits on Jan 3, 2019

• fix shared lib build

  Show description for 79f1a3b
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  79f1a3b

  Copy full SHA for 79f1a3b

Commits on Jan 2, 2019

• optimize remainder loops of requantization and rowoffset (pytorch#54)

  Show description for 172048a
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  172048a

  Copy full SHA for 172048a

• Fix a bug in FbgemmFP16 (pytorch#52)

  Show description for 4321c4d
  

  lifeiteng

  authored and

  facebook-github-bot

  committed
  4321c4d

  Copy full SHA for 4321c4d

• use 1 omp thread unless OMP_NUM_THREADS is explicitly set (pytorch#53)

  Show description for b86d642
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  b86d642

  Copy full SHA for b86d642

Commits on Dec 21, 2018

• Update the profiling format for Acc32 Benchmark (pytorch#50)

  Show description for bd35fce
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  bd35fce

  Copy full SHA for bd35fce

• Update with clang format (pytorch#51)

  Show description for 4691d5b
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  4691d5b

  Copy full SHA for 4691d5b

Commits on Dec 19, 2018

• Refactor to use FbgemmFP16 in packed gemm operator (pytorch#49)

  Show description for d5810be
  

  amylittleyang

  authored and

  facebook-github-bot

  committed
  d5810be

  Copy full SHA for d5810be

Commits on Dec 17, 2018

• add comments on col_offsets (pytorch#48)

  Show description for 1b3d970
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  1b3d970

  Copy full SHA for 1b3d970

Commits on Dec 11, 2018

• instantiate more kernels for PackAmatrix (pytorch#47)

  Show description for ebbe4f4
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  ebbe4f4

  Copy full SHA for ebbe4f4

Commits on Dec 6, 2018

• Fix duplicate symbols for thread local member variables (pytorch#43)

  Show description for 895646c
  

  James Reed

  authored and

  facebook-github-bot

  committed
  895646c

  Copy full SHA for 895646c

• Add missing <algorithm> include (pytorch#42)

  Show description for 8335a1a
  

  James Reed

  authored and

  facebook-github-bot

  committed
  8335a1a

  Copy full SHA for 8335a1a

Commits on Dec 5, 2018

• Final cleanup for avx2 isolation and consistent file names (pytorch#40)

  Show description for 3af8fe5
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  3af8fe5

  Copy full SHA for 3af8fe5

• avx2 intrinsic separation from OutputProcessing-inl.h (pytorch#38)

  Show description for 7e1b543
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  7e1b543

  Copy full SHA for 7e1b543

• Separate out avx2 code from dense x sparse matrix multiplication (pytorch#39)

  Show description for 32361c6
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  32361c6

  Copy full SHA for 32361c6

• File name change for FbgemmI8Depthwise.h and FbgemmI8Depthwise.cc (#14725)

  Show description for 8b3e0a7
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  8b3e0a7

  Copy full SHA for 8b3e0a7

• remove usage of c++ stdlib templates from FbgemmI8Depthwise (pytorch#37)

  Show description for d5ff1fa
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  d5ff1fa

  Copy full SHA for d5ff1fa

• clean up PackAWithQuantRowOffset from avx2 intrinsics (pytorch#36)

  Show description for a919889
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  a919889

  Copy full SHA for a919889

• Move avx2 intrinsics from PackAWithIm2Col (pytorch#35)

  Show description for f22e24d
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  f22e24d

  Copy full SHA for f22e24d

• Removed avx2 code from PackAWithRowOffset.cc (pytorch#34)

  Show description for b116758
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  b116758

  Copy full SHA for b116758

• avx2 specific code in a separate file for QuantUtils (pytorch#29)

  Show description for bf398ec
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  bf398ec

  Copy full SHA for bf398ec

• Move avx2 specific code in different source files (pytorch#28)

  Show description for 114a461
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  114a461

  Copy full SHA for 114a461

• Build status in readme on github

  Show description for 0d5a159
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  0d5a159

  Copy full SHA for 0d5a159

Commits on Dec 4, 2018

• Fix the group issue in the benchmark and use ResNext101 conv shapes (pytorch#32)

  Show description for 04b2e99
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  04b2e99

  Copy full SHA for 04b2e99

Commits on Dec 1, 2018

• Build fix with fbgemm shared lib (pytorch#31)

  Show description for 4fcc1f1
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  4fcc1f1

  Copy full SHA for 4fcc1f1

Commits on Nov 30, 2018

• Fix a bug in conv_ref

  Show description for 2fdb139
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  2fdb139

  Copy full SHA for 2fdb139

• Parallel build fails due to too much memory usage

  Show description for e1172a2
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  e1172a2

  Copy full SHA for e1172a2

• fixing CircleCI config file (pytorch#30)

  Show description for 1db75f6
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  1db75f6

  Copy full SHA for 1db75f6

• Only export symbols that are required while building shared library

  Show description for 4e9fa6c
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  4e9fa6c

  Copy full SHA for 4e9fa6c

• protect omp.h include by a pragma

  Show description for 4243ec0
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  4243ec0

  Copy full SHA for 4243ec0

Commits on Nov 29, 2018

• sparse convolution output processing (pytorch#27)

  Show description for 027de07
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  027de07

  Copy full SHA for 027de07

• Correct indentation

  Show description for 90535d3
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  90535d3

  Copy full SHA for 90535d3

Pagination

Previous

Next