Commits

Branch selector

 master

User selector

kimishpatel

Datepicker

All time

Commit History

Commits on Nov 19, 2020

• Introducing winograd transformed fp16 nnpack to PT for unet 106 (#47925)

  Show description for f0f8b97
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  f0f8b97

  Copy full SHA for f0f8b97

Commits on Nov 3, 2020

• For torch::jit::module's attr method to moble::module (#47059)

  Show description for 82b74bd
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  82b74bd

  Copy full SHA for 82b74bd

Commits on Oct 28, 2020

• Fix CPUCaching allocator guard bug (#46922)

  Show description for cbf90da
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  cbf90da

  Copy full SHA for cbf90da

Commits on Oct 15, 2020

• Bug fixes in profiling allocator (#45993)

  Show description for 4aaad88
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  4aaad88

  Copy full SHA for 4aaad88

Commits on Oct 6, 2020

• Disabling XNNPACK integration test in tsan mode (#45850)

  Show description for 001a799
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  001a799

  Copy full SHA for 001a799

• Profiling allocator for mobile. (#43951)

  Show description for a09e109
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  a09e109

  Copy full SHA for a09e109

Commits on Sep 29, 2020

• Move mobile specific CPUCachingAllocator to c10/mobile folder. (#45364)

  Show description for 6e55a26
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  6e55a26

  Copy full SHA for 6e55a26

Commits on Sep 17, 2020

• Implement better caching allocator for segmentation usecase. (#44618)

  Show description for f605d75
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  f605d75

  Copy full SHA for f605d75

Commits on Sep 3, 2020

• Fix replaceAtenConvolution for BC. (#44036)

  Show description for a153f69
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  a153f69

  Copy full SHA for a153f69

• Fix transposed conv2d rewrite pattern to account for convolution api (#44035)

  Show description for ba65cce
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  ba65cce

  Copy full SHA for ba65cce

Commits on Aug 28, 2020

• Fix bug in caching allocator. (#43719)

  Show description for dc5d365
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  dc5d365

  Copy full SHA for dc5d365

Commits on Aug 22, 2020

• Enable input pointer caching in XNNPACK integration. (#42840)

  Show description for b003f2c
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  b003f2c

  Copy full SHA for b003f2c

• Change quantizer to account for input tensor's memory format. (#42178)

  Show description for b52e6d0
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  b52e6d0

  Copy full SHA for b52e6d0

• Call qnnpack's conv setup only if input pointer has changed. (#42008)

  Show description for fb12992
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  fb12992

  Copy full SHA for fb12992

• Refactor qconv to reduce allocations. (#42007)

  Show description for 04aa42a
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  04aa42a

  Copy full SHA for 04aa42a

• Simple caching allocator for CPU. (#42006)

  Show description for 2a08566
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  2a08566

  Copy full SHA for 2a08566

Commits on Aug 17, 2020

• Fix freeze_module pass for sharedtype (#42457)

  Show description for 472f291
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  472f291

  Copy full SHA for 472f291

Commits on Aug 12, 2020

• Fix freeze_module pass for sharedtype (#42457)

  Show description for 4665f3f
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  4665f3f

  Copy full SHA for 4665f3f

Commits on Aug 5, 2020

• Revert D22939119: [TensorExpr] Fix a way we were createing np arrays in tests.

  Show description for 924a1db
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  924a1db

  Copy full SHA for 924a1db

Commits on Aug 3, 2020

• Calculate inverse of output scale first. (#41342)

  Show description for f0fd1cc
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  f0fd1cc

  Copy full SHA for f0fd1cc

Commits on Jul 24, 2020

• Support aarch32 neon backend for Vec256 (#41267)

  Show description for dede71d
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  dede71d

  Copy full SHA for dede71d

Commits on Jul 10, 2020

• Impilcit casting resulting internal build failure. (#41272)

  Show description for 1f2e91f
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  1f2e91f

  Copy full SHA for 1f2e91f

• Add add_relu fusion pass to optimize_for_mobile. (#40252)

  Show description for 8a79eec
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  8a79eec

  Copy full SHA for 8a79eec

Commits on Jul 9, 2020

• JIT pass for add relu fusion. (#39343)

  Show description for c5dcf05
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  c5dcf05

  Copy full SHA for c5dcf05

• Add fused add_relu op. (#39342)

  Show description for 82c9f79
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  82c9f79

  Copy full SHA for 82c9f79

• [Vec256][neon] Add neon backend for vec256 (#39341)

  Show description for d6feb61
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  d6feb61

  Copy full SHA for d6feb61

• Add benchmark for add op. (#40059)

  Show description for bddba1e
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  bddba1e

  Copy full SHA for bddba1e

Commits on Jun 29, 2020

• Add option to preserve certain methods during optimize_for_mobile. (#40629)

  Show description for 4a174c8
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  4a174c8

  Copy full SHA for 4a174c8

Commits on Jun 23, 2020

• Enabling concat fast path for channels last inputs (#39448)

  Show description for 6a421d5
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  6a421d5

  Copy full SHA for 6a421d5

Commits on Jun 19, 2020

• Run canonical graph optimizations in optimize_for_mobile. (#38840)

  Show description for ece8ef2
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  ece8ef2

  Copy full SHA for ece8ef2

Commits on Jun 18, 2020

• Move the check on orig_weight sizes. (#40200)

  Show description for 5f30950
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  5f30950

  Copy full SHA for 5f30950

• Add aarch64 specific quantize_tensor using arm intrinsics. (#40113)

  Show description for edd3fbc
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  edd3fbc

  Copy full SHA for edd3fbc

Commits on Jun 17, 2020

• Remove requirement of qnnpack engine for arm build. (#40112)

  Show description for 216f512
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  216f512

  Copy full SHA for 216f512

Commits on Jun 4, 2020

• Selective enabling of xnnpack based max_pool2d in ceil_mode. (#39447)

  Show description for eb5e037
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  eb5e037

  Copy full SHA for eb5e037

Commits on Jun 2, 2020

• Replace calls to contiguous with contiguous(suggested memory format) (#38433)

  Show description for 58cb369
  

  kimishpatel

  authored and

  facebook-github-bot

  committed
  58cb369

  Copy full SHA for 58cb369

Pagination

Previous

Next