Commits

Branch selector

 master

User selector

mrshenli

Datepicker

All time

Commit History

Commits on Feb 2, 2022

• Fix fx docs (pytorch#72108)

  Show description for 7c2eda3
  

  mrshenli

  authored and

  pytorchmergebot

  committed
  7c2eda3

  Copy full SHA for 7c2eda3

Commits on Jan 25, 2022

• Update distributed.rst for ProcessGroup Extensions (pytorch#71482)

  Show description for 7bc220e
  

  mrshenli

  authored and

  pytorchmergebot

  committed
  7bc220e

  Copy full SHA for 7bc220e

Commits on Nov 30, 2021

• Remove autograd-enabled collective APIs from distributed docs (pytorch#69011)

  Show description for b8c3693
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  b8c3693

  Copy full SHA for b8c3693

Commits on Nov 3, 2021

• Raise warning when calling collectives on non-member group objects (pytorch#67639)

  Show description for 18955d3
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  18955d3

  Copy full SHA for 18955d3

Commits on Oct 27, 2021

• Add barrier to ProcessGroup trampoline (pytorch#67236)

  Show description for a3aa9df
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  a3aa9df

  Copy full SHA for a3aa9df

• Enable c10d trampoline tests on MacOS (pytorch#67205)

  Show description for 632719c
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  632719c

  Copy full SHA for 632719c

Commits on Oct 26, 2021

• Prevent TCPServer get deleted too early (pytorch#67204)

  Show description for 1926156
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  1926156

  Copy full SHA for 1926156

• Consolidate c10d and dist imports in test_c10d_common.py (pytorch#67203)

  Show description for 2abffaf
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  2abffaf

  Copy full SHA for 2abffaf

Commits on Oct 21, 2021

• Setup c10d extension Backend class attr the same way as builtin ones (pytorch#66991)

  Show description for ce6f4b3
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  ce6f4b3

  Copy full SHA for ce6f4b3

Commits on Oct 11, 2021

• Add pybind trampoline for ProcessGroup and Work (pytorch#66338)

  Show description for 58fefa6
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  58fefa6

  Copy full SHA for 58fefa6

Commits on Sep 27, 2021

• Update ProcessGroup collective C++ APIs to be non-pure virtual functions (pytorch#64943)

  Show description for facff2e
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  facff2e

  Copy full SHA for facff2e

Commits on Sep 24, 2021

• Explicitly destory ProcessGroup in allgather_coalesced_async test (pytorch#65513)

  Show description for d78b390
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  d78b390

  Copy full SHA for d78b390

Commits on Sep 10, 2021

• Let all_reduce_coalesced and all_gather_coalesced return Future objects (pytorch#64722)

  Show description for 2a81e8b
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  2a81e8b

  Copy full SHA for 2a81e8b

Commits on Aug 27, 2021

• Improve distributed.get_rank() API docstring (pytorch#63296)

  Show description for 101a626
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  101a626

  Copy full SHA for 101a626

Commits on Aug 18, 2021

• Fix RPC Python User Function Error Handling (pytorch#63406)

  Show description for 3fd8e09
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  3fd8e09

  Copy full SHA for 3fd8e09

Commits on Aug 12, 2021

• Revert D30279364: [codemod][lint][fbcode/c*] Enable BLACK by default

  Show description for 1022443
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  1022443

  Copy full SHA for 1022443

Commits on Jun 17, 2021

• Run an dummy rpc._all_gather in init_rpc to avoid shutdown timeout (pytorch#59801)

  Show description for bbedfd9
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  bbedfd9

  Copy full SHA for bbedfd9

Commits on May 11, 2021

• Use RPC context streams to cover serde ops (pytorch#57926)

  Show description for cf7a0e5
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  cf7a0e5

  Copy full SHA for cf7a0e5

Commits on May 10, 2021

• Fix distributed autograd gradients synchronization (pytorch#57792)

  Show description for fc55290
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  fc55290

  Copy full SHA for fc55290

Commits on May 5, 2021

• Revert D28197820: [JIT][NNC] add hardswish symbolic gradient and NNC lowering

  Show description for 887d0e5
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  887d0e5

  Copy full SHA for 887d0e5

Commits on May 3, 2021

• Revert D28143091: [pytorch][PR] Add cross OpInfo

  Show description for 6bc3ad2
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  6bc3ad2

  Copy full SHA for 6bc3ad2

Commits on May 1, 2021

• Fix CUDA Stream synchronization when arguments contains RRefs (pytorch#57394)

  Show description for 2dffa8c
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  2dffa8c

  Copy full SHA for 2dffa8c

Commits on Apr 29, 2021

• Add devices argument to RRef constructor (pytorch#57085)

  Show description for 1ee54cc
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  1ee54cc

  Copy full SHA for 1ee54cc

Commits on Apr 27, 2021

• Use a FutureFactoryRegistry to allow libtorch_cpu files to create CUDAFuture (pytorch#56984)

  Show description for 5b3c0ae
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  5b3c0ae

  Copy full SHA for 5b3c0ae

• Remove unnecessary noCuda arg from AtomicJitFuture (pytorch#56973)

  Show description for f9e7e2e
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  f9e7e2e

  Copy full SHA for f9e7e2e

Commits on Apr 26, 2021

• Let RRef getValue() synchronize CUDA streams (pytorch#56895)

  Show description for d1088de
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  d1088de

  Copy full SHA for d1088de

Commits on Apr 23, 2021

• Add more RRef CUDA RPC tests (pytorch#56757)

  Show description for acca89e
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  acca89e

  Copy full SHA for acca89e

Commits on Apr 22, 2021

• Re-enable test_device_maps_gpu (pytorch#56415)

  Show description for b85b89d
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  b85b89d

  Copy full SHA for b85b89d

Commits on Apr 21, 2021

• Add devices to TensorPipe options (pytorch#56405)

  Show description for d83ae5d
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  d83ae5d

  Copy full SHA for d83ae5d

Commits on Apr 13, 2021

• Add a warning message to retire ProcessGroup RPC backend (pytorch#55616)

  Show description for e7bb00c
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  e7bb00c

  Copy full SHA for e7bb00c

Commits on Mar 18, 2021

• Avoid modifying rebuild buckets state in no_grad context (pytorch#54159)

  Show description for ef9ee46
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  ef9ee46

  Copy full SHA for ef9ee46

Commits on Mar 17, 2021

• Disallow CUDA RPC to use new devices in output tensors (pytorch#54024)

  Show description for e442d5c
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  e442d5c

  Copy full SHA for e442d5c

• Disable CUDA RPC tests that use new device in user-function outputs (pytorch#54023)

  Show description for 8cc06e3
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  8cc06e3

  Copy full SHA for 8cc06e3

Commits on Mar 8, 2021

• Fix set_device_map docs (pytorch#53508)

  Show description for 115df4f
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  115df4f

  Copy full SHA for 115df4f

• Use Store collect and verify names in all RPC agents (pytorch#53209)

  Show description for c7b1979
  

  mrshenli

  authored and

  facebook-github-bot

  committed
  c7b1979

  Copy full SHA for c7b1979

Pagination

Previous

Next