Commits

Branch selector

 master

User selector

desertfire

Datepicker

All time

Commit History

Commits on Aug 10, 2022

• [LTC] Pass a BackendDevice parameter into GetIrValueForScalarFromCodegen (pytorch#82970)

  Show description for 7b39406
  

  desertfire

  authored and

  pytorchmergebot

  committed
  7b39406

  Copy full SHA for 7b39406

Commits on May 26, 2022

• [LT] Add IR resuing support for manually-implemented ops

  Show description for 29189d2
  

  desertfire

  authored and

  pytorchmergebot

  committed
  29189d2

  Copy full SHA for 29189d2

Commits on May 16, 2022

• Revert "Revert "[LT] Codegen ReuseNode for supported ops""

  Show description for 25c6ebd
  

  desertfire

  authored and

  pytorchmergebot

  committed
  25c6ebd

  Copy full SHA for 25c6ebd

Commits on May 13, 2022

• [LT] Codegen ReuseNode for supported ops

  Show description for 6066e59
  

  desertfire

  authored and

  pytorchmergebot

  committed
  6066e59

  Copy full SHA for 6066e59

Commits on May 11, 2022

• Revert "Revert "[LT] Store OpKind for each IR subclass in a static field""

  Show description for 8f5cdc6
  

  desertfire

  authored and

  pytorchmergebot

  committed
  8f5cdc6

  Copy full SHA for 8f5cdc6

Commits on May 9, 2022

• [LT] Move device lock in LazyGraphExecutor to a later place

  Show description for 36150c6
  

  desertfire

  authored and

  pytorchmergebot

  committed
  36150c6

  Copy full SHA for 36150c6

Commits on May 6, 2022

• [LT] Store OpKind for each IR subclass in a static field

  Show description for ac37ddc
  

  desertfire

  authored and

  pytorchmergebot

  committed
  ac37ddc

  Copy full SHA for ac37ddc

Commits on May 4, 2022

• [LT] Add a trie data structure for caching IR nodes

  Show description for f05710d
  

  desertfire

  authored and

  pytorchmergebot

  committed
  f05710d

  Copy full SHA for f05710d

Commits on May 3, 2022

• [LT] Move MakeNode into ir_builder.h

  Show description for f8a4780
  

  desertfire

  authored and

  pytorchmergebot

  committed
  f8a4780

  Copy full SHA for f8a4780

• [LT] Add a flag to control IR reusing

  Show description for 65b9778
  

  desertfire

  authored and

  pytorchmergebot

  committed
  65b9778

  Copy full SHA for 65b9778

Commits on Jan 7, 2022

• Dispatch to at::convolution intead of at::_convolution in _convolution_double_backward (pytorch#70661)

  Show description for f135438
  

  desertfire

  authored and

  facebook-github-bot

  committed
  f135438

  Copy full SHA for f135438

Commits on Jan 5, 2022

• [LTC] Fix stride accessors in LTCTensorImpl (pytorch#70623)

  Show description for 70d3b27
  

  desertfire

  authored and

  facebook-github-bot

  committed
  70d3b27

  Copy full SHA for 70d3b27

Commits on Dec 16, 2021

• [LTC] Upstream LazyTensor and LazyGraphExecutor (pytorch#69815)

  Show description for fe7b644
  

  desertfire

  authored and

  facebook-github-bot

  committed
  fe7b644

  Copy full SHA for fe7b644

• [LTC] Upstream several internal ops (pytorch#69716)

  Show description for 2824376
  

  desertfire

  authored and

  facebook-github-bot

  committed
  2824376

  Copy full SHA for 2824376

• [LTC] Upstream utils in computation_client (pytorch#69621)

  Show description for e6a4988
  

  desertfire

  authored and

  facebook-github-bot

  committed
  e6a4988

  Copy full SHA for e6a4988

Commits on Dec 8, 2021

• [LT] Sync with the lazy_tensor_staging branch (pytorch#69527)

  Show description for 8a975c0
  

  desertfire

  authored and

  facebook-github-bot

  committed
  8a975c0

  Copy full SHA for 8a975c0

Commits on Dec 4, 2021

• [LT] Upstream LazyView and view ops IR Nodes (pytorch#69277)

  Show description for e8f4c9c
  

  desertfire

  authored and

  facebook-github-bot

  committed
  e8f4c9c

  Copy full SHA for e8f4c9c

• [LT] Upstream more util functions (pytorch#69098)

  Show description for 0bbe21b
  

  desertfire

  authored and

  facebook-github-bot

  committed
  0bbe21b

  Copy full SHA for 0bbe21b

Commits on Nov 30, 2021

• [LT] Sync LTC branch changes on torch/csrc/lazy/core (pytorch#69012)

  Show description for b83e8d5
  

  desertfire

  authored and

  facebook-github-bot

  committed
  b83e8d5

  Copy full SHA for b83e8d5

Commits on Nov 29, 2021

• Add lazy::Shape::numel() (pytorch#68314)

  Show description for 787ded5
  

  desertfire

  authored and

  facebook-github-bot

  committed
  787ded5

  Copy full SHA for 787ded5

Commits on Nov 9, 2021

• [LT] Merge cache.h (pytorch#67929)

  Show description for a027551
  

  desertfire

  authored and

  facebook-github-bot

  committed
  a027551

  Copy full SHA for a027551

• [LT] Merge permutation_util into master (pytorch#67766)

  Show description for a473417
  

  desertfire

  authored and

  facebook-github-bot

  committed
  a473417

  Copy full SHA for a473417

Commits on Oct 28, 2021

• [LT] Add ir_util for ComputePostOrder (pytorch#67282)

  Show description for 2366948
  

  desertfire

  authored and

  facebook-github-bot

  committed
  2366948

  Copy full SHA for 2366948

Commits on Oct 5, 2021

• [JIT] Initialize CUDA context before launching fused kernel (pytorch#65064)

  Show description for 6e06cb7
  

  desertfire

  authored and

  facebook-github-bot

  committed
  6e06cb7

  Copy full SHA for 6e06cb7

Commits on Jun 17, 2021

• [NNC] Handle int64 indices and loop bounds (pytorch#59769)

  Show description for 3dc8112
  

  desertfire

  authored and

  facebook-github-bot

  committed
  3dc8112

  Copy full SHA for 3dc8112

• [NNC] Add a dtypeToCppString virtual method in IRPrinter (pytorch#59449)

  Show description for 96b3537
  

  desertfire

  authored and

  facebook-github-bot

  committed
  96b3537

  Copy full SHA for 96b3537

Commits on Jun 3, 2021

• [JIT] Add a phase to perform inplace<->functional conversion for activation operators (pytorch#57477)

  Show description for add291c
  

  desertfire

  authored and

  facebook-github-bot

  committed
  add291c

  Copy full SHA for add291c

Commits on May 27, 2021

• Enable NNC fusion for relu6 (pytorch#58773)

  Show description for 7e4e648
  

  desertfire

  authored and

  facebook-github-bot

  committed
  7e4e648

  Copy full SHA for 7e4e648

Commits on May 6, 2021

• [JIT] Lazily initialize aliasDb in DCE (pytorch#56649)

  Show description for eb6445a
  

  desertfire

  authored and

  facebook-github-bot

  committed
  eb6445a

  Copy full SHA for eb6445a

Commits on Apr 30, 2021

• [JIT] Enable conv-add-relu fusion as a part of frozen graph optimization (pytorch#56580)

  Show description for c1a4422
  

  desertfire

  authored and

  facebook-github-bot

  committed
  c1a4422

  Copy full SHA for c1a4422

• [JIT] Disable conv-add-relu fusion for cuDNN7 when model uses fp16 (pytorch#56579)

  Show description for c971401
  

  desertfire

  authored and

  facebook-github-bot

  committed
  c971401

  Copy full SHA for c971401

Commits on Apr 12, 2021

• [JIT] Add cond-add-relu matching pattern to cover in-place ops (pytorch#55458)

  Show description for 69b7b01
  

  desertfire

  authored and

  facebook-github-bot

  committed
  69b7b01

  Copy full SHA for 69b7b01

Commits on Apr 6, 2021

• [JIT] Include conv3d in the conv-add-relu fusion (pytorch#54772)

  Show description for a0d9776
  

  desertfire

  authored and

  facebook-github-bot

  committed
  a0d9776

  Copy full SHA for a0d9776

Commits on Mar 18, 2021

• [JIT] Add CUDNN Conv-Add-Relu fusion for Frozen Model Optimization (pytorch#52102)

  Show description for 4626886
  

  desertfire

  authored and

  facebook-github-bot

  committed
  4626886

  Copy full SHA for 4626886

Commits on Feb 8, 2021

• [JIT] Remove Dropout during Frozon Optimization (pytorch#51589)

  Show description for 0dd1d60
  

  desertfire

  authored and

  facebook-github-bot

  committed
  0dd1d60

  Copy full SHA for 0dd1d60