Commits

Branch selector

 master

User selector

apaszke

All time

Commit History

Commits on Feb 15, 2019

• Add special ops for BatchNorm symbolic differentiation (pytorch#15403)

  

  apaszke

  authored and

  facebook-github-bot

  committedFeb 15, 2019
  7157be8

Commits on Feb 11, 2019

• Don't keep unnecessary saved_inputs alive (pytorch#16583)

  

  apaszke

  authored and

  facebook-github-bot

  committedFeb 11, 2019
  7743ed8

• Add support for fusion of half batch norm with float stats (pytorch#16735)

  

  apaszke

  authored and

  facebook-github-bot

  committedFeb 11, 2019
  8f6ee88

Commits on Feb 5, 2019

• Make tuple checks faster (pytorch#16657)

  

  apaszke

  authored and

  facebook-github-bot

  committedFeb 5, 2019
  963e410

• Don't throw in operator== for TypeMeta and ScalarType (pytorch#16736)

  

  apaszke

  authored and

  facebook-github-bot

  committedFeb 5, 2019
  c751cf8

Commits on Jan 11, 2019

• Simplify cat fusion (pytorch#15633)

  

  apaszke

  authored and

  facebook-github-bot

  committedJan 11, 2019
  d580d35

Commits on Jan 10, 2019

• JIT Batch Norm fusion (pytorch#15897)

  

  apaszke

  authored and

  facebook-github-bot

  committedJan 10, 2019
  d35295c

Commits on Jan 8, 2019

• Stop leaving garbage files after running test_jit.py

  

  apaszke

  authored and

  facebook-github-bot

  committedJan 8, 2019
  29a9d6a

• Add support for batch_norm fusion to the JIT (pytorch#15146)

  

  apaszke

  authored and

  facebook-github-bot

  committedJan 8, 2019
  5e1b35b

Commits on Dec 5, 2018

• Use AT_WARN for warnings in the JIT (pytorch#14770)

  

  apaszke

  authored and

  facebook-github-bot

  committedDec 5, 2018
  f89de64

• Don't DCE PythonOp

  

  apaszke

  authored and

  facebook-github-bot

  committedDec 5, 2018
  c79e305

• Improvements for symbolic AD (pytorch#14758)

  

  apaszke

  authored and

  facebook-github-bot

  committedDec 5, 2018
  8dfebc1

Commits on Dec 4, 2018

• Reenable all forward-pass fusions that worked before the AD fix (pytorch#14558)

  

  apaszke

  authored and

  facebook-github-bot

  committedDec 4, 2018
  d76fd43

• Disable randn_like fusion in the JIT (pytorch#14752)

  

  apaszke

  authored and

  facebook-github-bot

  committedDec 4, 2018
  7bc489c

• Reduce broadcasted inputs in derivative code (pytorch#14485)

  

  apaszke

  authored and

  facebook-github-bot

  committedDec 4, 2018
  8812a5d

Commits on Dec 3, 2018

• Reduce broadcasted inputs in derivative code (pytorch#14485)

  

  apaszke

  authored and

  facebook-github-bot

  committedDec 3, 2018
  68ffe46

Commits on Nov 29, 2018

• Broadcast prim::FusedConcat inputs independently when checking kernels (pytorch#14503)

  

  apaszke

  authored and

  facebook-github-bot

  committedNov 29, 2018
  31b3d81

Commits on Nov 26, 2018

• Batch more matrix multiplies (pytorch#13456)

  

  apaszke

  authored and

  facebook-github-bot

  committedNov 26, 2018
  a603689

Commits on Nov 6, 2018

• Append parameters when checking graphs for TorchScript Methods (pytorch#13553)

  

  apaszke

  authored and

  facebook-github-bot

  committedNov 6, 2018
  f6ff5d8

Commits on Nov 5, 2018

• Stop depending on static analysis of tensor types in graph fuser (pytorch#13387)

  

  apaszke

  authored and

  facebook-github-bot

  committedNov 5, 2018
  e988dc6

Commits on Nov 1, 2018

• Re-enabled mm+add tree batching in the JIT (pytorch#13228)

  

  apaszke

  authored and

  facebook-github-bot

  committedNov 1, 2018
  26a8bb6

Commits on Sep 26, 2018

• Fix ONNX bug, add symbolic for full

  

  apaszke

  authored and

  facebook-github-bot

  committedSep 26, 2018
  78fe149

• Enable tracing of tensor factories with an out argument

  

  apaszke

  authored and

  facebook-github-bot

  committedSep 26, 2018
  18f9c07

Commits on Sep 25, 2018

• Eliminate no-op adds and muls in peephole pass (pytorch#11801)

  

  apaszke

  authored and

  facebook-github-bot

  committedSep 25, 2018
  a830964

Commits on Sep 24, 2018

• Stop moving constants into DifferentiableSubgraphs (pytorch#11809)

  

  apaszke

  authored and

  facebook-github-bot

  committedSep 24, 2018
  5141482

Commits on Sep 21, 2018

• Specialize ArgumentSpecs on tuple elements too (pytorch#11863)

  

  apaszke

  authored and

  facebook-github-bot

  committedSep 21, 2018
  7efbf3a

• Minor JIT improvements (pytorch#11654)

  

  apaszke

  authored and

  facebook-github-bot

  committedSep 21, 2018
  1ad7e0c

Commits on Sep 19, 2018

• Improve autograd profiler performance (pytorch#11773)

  

  apaszke

  authored and

  facebook-github-bot

  committedSep 19, 2018
  8c3a94e

Commits on Sep 14, 2018

• Implement requires_grad propagation in the JIT (pytorch#11586)

  

  apaszke

  authored and

  facebook-github-bot

  committedSep 14, 2018
  98e04db

Commits on Sep 12, 2018

• Improve tracer warnings (pytorch#11545)

  

  apaszke

  authored and

  facebook-github-bot

  committedSep 12, 2018
  90e31f4

• Make .to() methods native functions (to fix JIT tracing)

  

  apaszke

  authored and

  facebook-github-bot

  committedSep 12, 2018
  62c9d4a

• Release GIL when calling into JIT interpreter

  

  apaszke

  authored and

  facebook-github-bot

  committedSep 12, 2018
  a00fa2c

• Allow tracing random functions (only when using default generators) (pytorch#11539)

  

  apaszke

  authored and

  facebook-github-bot

  committedSep 12, 2018
  8b196d6

Commits on Sep 11, 2018

• Remove time prefix from rsync (pytorch#11525)

  

  apaszke

  authored and

  facebook-github-bot

  committedSep 11, 2018
  781737f

• Add support for tracing strings (pytorch#11506)

  

  apaszke

  authored and

  facebook-github-bot

  committedSep 11, 2018
  120d769

Pagination

PreviousNext