Commits

Branch selector

 master

User selector

apaszke

Datepicker

All time

Commit History

Commits on Nov 2, 2017

• Update log_softmax and softmax signatures to include dim (pytorch#106)

  

  apaszke

  authored and

  ezyang

  committed
  8257f3e

  Copy full SHA for 8257f3e

Commits on Nov 1, 2017

• Improve NYI error message to point to VariableType

  apaszke

  committed
  53b0152

  Copy full SHA for 53b0152

• Register VariableType methods in autograd profiler

  apaszke

  committed
  6e17e73

  Copy full SHA for 6e17e73

Commits on Oct 28, 2017

• Replace Variable constructor with a static_cast in tracer

  apaszke

  committed
  6d2e395

  Copy full SHA for 6d2e395

• Add a hack for RNN export to ONNX

  apaszke

  committed
  a381fa1

  Copy full SHA for a381fa1

Commits on Oct 26, 2017

• Re-enable and fix most JIT tests

  

  apaszke

  authored and

  soumith

  committed
  fa0f3cf

  Copy full SHA for fa0f3cf

• Autogenerate ATen dispatch for JIT nodes

  

  apaszke

  authored and

  soumith

  committed
  61afb0d

  Copy full SHA for 61afb0d

• Fix pack_padded_sequence to accept inputs of arbitrary sizes

  apaszke

  committed
  5760b03

  Copy full SHA for 5760b03

Commits on Oct 25, 2017

• Softmax/LogSoftMax refactor (wrapped up) (pytorch#3245)

  Show description for b3642b3
  

  apaszke

  authored and

  soumith

  committed
  b3642b3

  Copy full SHA for b3642b3

Commits on Oct 20, 2017

• Fix nvprof mode in autograd profiler

  

  apaszke

  authored and

  ezyang

  committed
  76abc06

  Copy full SHA for 76abc06

Commits on Oct 19, 2017

• Large Softmax and LogSoftmax refactor

  Show description for 98e6744
  apaszke

  committed
  98e6744

  Copy full SHA for 98e6744

• Add support for saving the output in autogenerated functions

  apaszke

  committed
  3a4ca7a

  Copy full SHA for 3a4ca7a

• Update DLPack bindings

  apaszke

  committed
  f9ee52e

  Copy full SHA for f9ee52e

• Merge commit '99cbf24b8b5a9d769b9794e447e0b740bcdd99c8'

  apaszke

  committed
  76071cf

  Copy full SHA for 76071cf

• Update log_softmax and softmax signatures to include dim (pytorch#106)

  

  apaszke

  authored and

  soumith

  committed
  99cbf24

  Copy full SHA for 99cbf24

Commits on Oct 11, 2017

• Fix macOS build (with CUDA) (pytorch#3071)

  

  apaszke

  authored and

  soumith

  committed
  cc3058b

  Copy full SHA for cc3058b

Commits on Oct 3, 2017

• Add inplace checks in JIT

  

  apaszke

  authored and

  soumith

  committed
  b6b41c8

  Copy full SHA for b6b41c8

• Fix THC exponential to not sample infinity

  

  apaszke

  authored and

  soumith

  committed
  82bc97e

  Copy full SHA for 82bc97e

• Add CUDNN_INCLUDE_DIR before CUDA directories in setup.py

  

  apaszke

  authored and

  soumith

  committed
  437d3af

  Copy full SHA for 437d3af

• Hotpatch THPP compile error

  

  apaszke

  authored and

  soumith

  committed
  bf82ecd

  Copy full SHA for bf82ecd

• Limit number of demangler invocations in autograd profiler

  

  apaszke

  authored and

  soumith

  committed
  6fbbb1b

  Copy full SHA for 6fbbb1b

Commits on Sep 30, 2017

• Simplify getApplyGrid in THC (pytorch#2900)

  

  apaszke

  authored and

  soumith

  committed
  9b9704e

  Copy full SHA for 9b9704e

Commits on Sep 26, 2017

• Add CUDA profiler bindings

  

  apaszke

  authored and

  soumith

  committed
  833bedc

  Copy full SHA for 833bedc

• Always regenerate nn wrappers after rebuilding THNN and THCUNN

  

  apaszke

  authored and

  soumith

  committed
  b784966

  Copy full SHA for b784966

• Add tools for autograd profiling

  

  apaszke

  authored and

  soumith

  committed
  411e146

  Copy full SHA for 411e146

Commits on Sep 25, 2017

• Allow specifying unused inputs to torch.autograd.grad (pytorch#2859)

  

  apaszke

  authored and

  soumith

  committed
  3128218

  Copy full SHA for 3128218

• Make distributed recv return sender rank

  

  apaszke

  authored and

  soumith

  committed
  2a8603c

  Copy full SHA for 2a8603c

Commits on Sep 20, 2017

• Minor rebase fixes

  apaszke

  committed
  16a3de0

  Copy full SHA for 16a3de0

Commits on Sep 19, 2017

• Remove TensorMeta

  apaszke

  committed
  c536da7

  Copy full SHA for c536da7

• Prune null edges in Eval nodes

  apaszke

  committed
  a7c4152

  Copy full SHA for a7c4152

• Add a pass to remove all non-standard ONNX nodes before export (pytorch#225)

  apaszke

  committed
  b66d90c

  Copy full SHA for b66d90c

• Add support for some ONNX nodes in JIT closure

  apaszke

  committed
  b7e89d7

  Copy full SHA for b7e89d7

• Handle AddConstant in fusion compiler

  apaszke

  committed
  fe5c644

  Copy full SHA for fe5c644

• Make sure passes don't mess up stages of nodes and graphs

  apaszke

  committed
  e05cfb2

  Copy full SHA for e05cfb2

• Minor refactor of fusion compiler

  apaszke

  committed
  8a605ce

  Copy full SHA for 8a605ce

Pagination

Previous

Next