Commits

Branch selector

 tensor-merge

User selector

All users

Datepicker

All time

Commit History

Commits on Jul 24, 2018

• Add missing windows exports

  Show description for 9bf72b2
  

  zdevito

  authored and

  facebook-github-bot

  committed
  9bf72b2

  Copy full SHA for 9bf72b2

• Add 1x1 specialization for conv with NCHW order (#9671)

  Show description for 5df3eae
  

  xiaomengy

  authored and

  facebook-github-bot

  committed
  5df3eae

  Copy full SHA for 5df3eae

• Re-enable test_segfault after recent dataloder changes

  Show description for a387331
  

  ssnl

  authored and

  facebook-github-bot

  committed
  a387331

  Copy full SHA for a387331

• Tensor merge PRs from July 20 (#9713)

  Show description for 099b5ba
  

  ezyang

  authored and

  facebook-github-bot

  committed
  099b5ba

  Copy full SHA for 099b5ba

Commits on Jul 23, 2018

• Allow multiple ops.def and clean up code gen in general

  Show description for e3fb908
  

  bwasti

  authored and

  facebook-github-bot

  committed
  e3fb908

  Copy full SHA for e3fb908

• Add operator<< overloads for TensorOptions (#9606)

  Show description for 5849354
  

  goldsborough

  authored and

  facebook-github-bot

  committed
  5849354

  Copy full SHA for 5849354

• Change behavior of clone to clone to a device (#9609)

  Show description for d05a814
  

  goldsborough

  authored and

  facebook-github-bot

  committed
  d05a814

  Copy full SHA for d05a814

• Rename embedding variable to weight (#9720)

  Show description for 31ba2f1
  

  goldsborough

  authored and

  facebook-github-bot

  committed
  31ba2f1

  Copy full SHA for 31ba2f1

• quick patch for PackPadded removal to propagate the correct size. (#9657)

  Show description for 431415a
  

  anderspapitto

  authored and

  facebook-github-bot

  committed
  431415a

  Copy full SHA for 431415a

• Switch interpreter to use IValue's primitive int/floats (#9718)

  Show description for a949245
  

  zdevito

  authored and

  facebook-github-bot

  committed
  a949245

  Copy full SHA for a949245

• Add fallback to TensorCPU if there are unsupported types for IDEEP Tensor (#9667)

  Show description for a9742e1
  

  Yinghai Lu

  authored and

  facebook-github-bot

  committed
  a9742e1

  Copy full SHA for a9742e1

• Add ctc_beam_search_decoder op for caffe2 (#9622)

  Show description for ee2cc68
  

  normster

  authored and

  facebook-github-bot

  committed
  ee2cc68

  Copy full SHA for ee2cc68

• Extend DispatchStub to support CUDA dispatch (#9664)

  Show description for aa8a9fa
  

  colesbury

  authored and

  facebook-github-bot

  committed
  aa8a9fa

  Copy full SHA for aa8a9fa

• Improving diagnose RF NE with Cali (#9550)

  Show description for 3e9e3ef
  

  Xiaolong Wang

  authored and

  facebook-github-bot

  committed
  3e9e3ef

  Copy full SHA for 3e9e3ef

• Fix D8722560 (#9717)

  Show description for 88d6b6e
  

  smessmer

  authored and

  facebook-github-bot

  committed
  88d6b6e

  Copy full SHA for 88d6b6e

• Create torch::from_blob for variables (#9605)

  Show description for 5094684
  

  goldsborough

  authored and

  facebook-github-bot

  committed
  5094684

  Copy full SHA for 5094684

• Reformat output data format to make it more general for other binaries (#9555)

  Show description for 14d4bdb
  

  sf-wind

  authored and

  facebook-github-bot

  committed
  14d4bdb

  Copy full SHA for 14d4bdb

• Improve error messages of wrong dimensions (#9694)

  Show description for 029cf1d
  

  idansc

  authored and

  facebook-github-bot

  committed
  029cf1d

  Copy full SHA for 029cf1d

• Low rank multivariate normal (#8635)

  Show description for 9525925
  

  fehiepsi

  authored and

  facebook-github-bot

  committed
  9525925

  Copy full SHA for 9525925

• Support n-dimensional empty tensors in CUDA non-reduction dimension f… (#9658)

  Show description for 9d6521c
  

  gchanan

  authored and

  facebook-github-bot

  committed
  9d6521c

  Copy full SHA for 9d6521c

• Remove CMAKE_WINDOWS_EXPORT_ALL_SYMBOLS and fix CUDA 8 build on Windows (#9491) (#9491)

  Show description for 53083b8
  

  peterjc123

  authored and

  facebook-github-bot

  committed
  53083b8

  Copy full SHA for 53083b8

• Fix dataloader hang when it is not completely iterated (#9655)

  Show description for 9ee5133
  

  ssnl

  authored and

  facebook-github-bot

  committed
  9ee5133

  Copy full SHA for 9ee5133

Commits on Jul 22, 2018

• Hide all other fields in THTensor (#9683)

  Show description for 1afdc57
  

  ezyang

  authored and

  facebook-github-bot

  committed
  1afdc57

  Copy full SHA for 1afdc57

• Modify barrier net to allow better control over its initialization and execution in DPM (#9665)

  Show description for f3d72b2
  

  ygdx

  authored and

  facebook-github-bot

  committed
  f3d72b2

  Copy full SHA for f3d72b2

• Add new ways of matching nodes with schemas in the JIT (#9567)

  Show description for 769cb5a
  

  apaszke

  authored and

  facebook-github-bot

  committed
  769cb5a

  Copy full SHA for 769cb5a

Commits on Jul 21, 2018

• Update channel_shuffle_op and transpose 2d to speed up ShuffleNet (#9525)

  Show description for a01d6f0
  

  xiaomengy

  authored and

  facebook-github-bot

  committed
  a01d6f0

  Copy full SHA for a01d6f0

• Allow MKLDNN on macOS, and any other OS where CMake is able to detect it.

  Show description for 3bb8c5e
  

  Owen Anderson

  authored and

  facebook-github-bot

  committed
  3bb8c5e

  Copy full SHA for 3bb8c5e

• Add a CUDAContext header include

  Show description for b5c8d59
  

  ezyang

  authored and

  facebook-github-bot

  committed
  b5c8d59

  Copy full SHA for b5c8d59

• Guard include of cuda-only header comm.h (#9656)

  Show description for 23ed26a
  

  ezyang

  authored and

  facebook-github-bot

  committed
  23ed26a

  Copy full SHA for 23ed26a

• Fix for half conversion for ROCm 1.8.2 (#9663)

  Show description for 5e84403
  

  Ashish

  authored and

  facebook-github-bot

  committed
  5e84403

  Copy full SHA for 5e84403

Commits on Jul 20, 2018

• Support n-dimensional empty tensors in take/put.

  Show description for 3efdece
  

  gchanan

  authored and

  facebook-github-bot

  committed
  3efdece

  Copy full SHA for 3efdece

• ONNXIFI transform (#9569)

  Show description for 45e5c17
  

  Yinghai Lu

  authored and

  facebook-github-bot

  committed
  45e5c17

  Copy full SHA for 45e5c17

• Add workspace.RunPlanInBackground (#9637)

  Show description for 0158103
  

  kittipatv

  authored and

  facebook-github-bot

  committed
  0158103

  Copy full SHA for 0158103

• Creates CUDAContext (#9435)

  Show description for 1003ccf
  

  mruberry

  authored and

  facebook-github-bot

  committed
  1003ccf

  Copy full SHA for 1003ccf

• set_input_record() should always add external input (#9636)

  Show description for 8a0fe0a
  

  kittipatv

  authored and

  facebook-github-bot

  committed
  8a0fe0a

  Copy full SHA for 8a0fe0a

Pagination

Previous

Next