Commits

Branch selector

 master

User selector

dskhudia

Datepicker

All time

Commit History

Commits on Aug 30, 2021

• Back out "[JIT] Add aten::slice optimization"

  Show description for 65050ec
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  65050ec

  Copy full SHA for 65050ec

Commits on Oct 21, 2020

• [caffe2][dnnlowp] Remove openmp usage in quantize dnnlowp op

  Show description for f47231b
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  f47231b

  Copy full SHA for f47231b

Commits on Oct 12, 2020

• [pt][quant] Support either min or max in qclamp (pytorch#45937)

  Show description for 87a4baf
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  87a4baf

  Copy full SHA for 87a4baf

Commits on Sep 25, 2020

• [aten] Call fbgemm functions for embedding prepack/unpack (pytorch#44845)

  Show description for 677a59d
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  677a59d

  Copy full SHA for 677a59d

Commits on Sep 22, 2020

• [caffe2] Replace embedding conversion ops with fbgemm functions (pytorch#44843)

  Show description for 09aee06
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  09aee06

  Copy full SHA for 09aee06

Commits on Sep 3, 2020

• [fbgemm] manual submodule update (pytorch#44082)

  Show description for 7d95eb8
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  7d95eb8

  Copy full SHA for 7d95eb8

Commits on May 1, 2020

• [pt][quant] Unify numerics between fakequant and quant/dequant (pytorch#37188)

  Show description for fa8ab4b
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  fa8ab4b

  Copy full SHA for fa8ab4b

Commits on Apr 15, 2020

• [pt][quant] avgpool3d for graph mode (pytorch#36598)

  Show description for df9a250
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  df9a250

  Copy full SHA for df9a250

Commits on Apr 9, 2020

• [pt][quant] Add vector path to copy kernel for quantized data types (pytorch#36189)

  Show description for 866227c
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  866227c

  Copy full SHA for 866227c

Commits on Mar 23, 2020

• [pt][quant] Optimized qadd_scalar (pytorch#34925)

  Show description for 506996c
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  506996c

  Copy full SHA for 506996c

Commits on Mar 19, 2020

• [pt][quant] qmul and qadd should preserve input memory format (pytorch#34834)

  Show description for 7335f07
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  7335f07

  Copy full SHA for 7335f07

Commits on Mar 10, 2020

• [pt][quant] Vectorized qmul and more methods on qint data types (pytorch#34376)

  Show description for a09c4d3
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  a09c4d3

  Copy full SHA for a09c4d3

Commits on Feb 26, 2020

• [pt][quant] Parallelize quantize and dequantize (pytorch#33765)

  Show description for a8e7ed4
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  a8e7ed4

  Copy full SHA for a8e7ed4

Commits on Dec 12, 2019

• Adding quantized clamp kernel (pytorch#30541)

  Show description for a2463cb
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  a2463cb

  Copy full SHA for a2463cb

Commits on Nov 27, 2019

• Fix docs so that the example works (pytorch#30120)

  Show description for 2d6b2f3
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  2d6b2f3

  Copy full SHA for 2d6b2f3

Commits on Nov 19, 2019

• Build time warning on windows for fbgemm (pytorch#29062)

  Show description for 79b797c
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  79b797c

  Copy full SHA for 79b797c

Commits on Oct 2, 2019

• Enabling intra-op parallelism (pytorch#26692)

  Show description for 3eefc54
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  3eefc54

  Copy full SHA for 3eefc54

Commits on Oct 1, 2019

• Set number of threads for operator_benchmarks (pytorch#27010)

  Show description for bf7ebc5
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  bf7ebc5

  Copy full SHA for bf7ebc5

Commits on Sep 24, 2019

• Remove duplicate calculation of output shape (pytorch#26684)

  Show description for fa7b621
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  fa7b621

  Copy full SHA for fa7b621

Commits on Sep 23, 2019

• fix operator level benchmark to have NHWC layout (pytorch#26577)

  Show description for fc926d9
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  fc926d9

  Copy full SHA for fc926d9

Commits on Sep 17, 2019

• Enable support for dilated convolutions (pytorch#26205)

  Show description for 0ad8c67
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  0ad8c67

  Copy full SHA for 0ad8c67

Commits on Sep 16, 2019

• Dynamic quantization for bias. (pytorch#26057)

  Show description for 2b52c1d
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  2b52c1d

  Copy full SHA for 2b52c1d

Commits on Sep 11, 2019

• Improve error message when input is not in the right format (pytorch#25928)

  Show description for ec48280
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  ec48280

  Copy full SHA for ec48280

Commits on Aug 30, 2019

• Minor fixes in per channel support for qconv kernel (pytorch#25182)

  Show description for fcab254
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  fcab254

  Copy full SHA for fcab254

Commits on Aug 26, 2019

• per channel quantization support (pytorch#25134)

  Show description for d806257
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  d806257

  Copy full SHA for d806257

Commits on Aug 23, 2019

• Add missing functions and methods for channelwise quantization (pytorch#24934)

  Show description for 12ea1d7
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  12ea1d7

  Copy full SHA for 12ea1d7

• per channel quantization support (pytorch#24936)

  Show description for 9e99650
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  9e99650

  Copy full SHA for 9e99650

• Per Channel quantization APIs (pytorch#24935)

  Show description for aba15ce
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  aba15ce

  Copy full SHA for aba15ce

Commits on Aug 13, 2019

• Enable FBGEMM tests under UBSAN as well (pytorch#23570)

  Show description for f510409
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  f510409

  Copy full SHA for f510409

Commits on Aug 9, 2019

• Fix qconv benchmark (pytorch#24019)

  Show description for aa02b1a
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  aa02b1a

  Copy full SHA for aa02b1a

• fix scale and zero_point names (pytorch#23991)

  Show description for a055678
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  a055678

  Copy full SHA for a055678

Commits on Aug 5, 2019

• qconv operator level benchmark (pytorch#22895)

  Show description for fb06c9e
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  fb06c9e

  Copy full SHA for fb06c9e

Commits on Aug 4, 2019

• Support for non-zero zero_points for weight and activation (pytorch#23541)

  Show description for 3113773
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  3113773

  Copy full SHA for 3113773

Commits on Jul 30, 2019

• Update argument list for non-fbgemm path for qconv_prepack (pytorch#23521)

  Show description for bfee46f
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  bfee46f

  Copy full SHA for bfee46f

Commits on Jul 26, 2019

• qconv+relu and qlinear+relu modules (pytorch#23410)

  Show description for 4104e80
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  4104e80

  Copy full SHA for 4104e80

Pagination

Previous

Next