Commits

Branch selector

 ncclfix

User selector

All users

Datepicker

All time

Commit History

Commits on Apr 29, 2017

• fix NCCL makefile for CUDA 7.5

  soumith

  committed
  27fb875

  Copy full SHA for 27fb875

Commits on Apr 28, 2017

• remove inplace pow and fix contiguous -> coalesce (#1398)

  soumith

  authored
  45020a7

  Copy full SHA for 45020a7

• Document hybrid sparse tensors.

  Show description for 9c01f5d
  

  ezyang

  authored and

  apaszke

  committed
  9c01f5d

  Copy full SHA for 9c01f5d

• Add new init methods gain, eye and dirac (#1172)

  

  Kaixhin

  authored and

  soumith

  committed
  cbb9f08

  Copy full SHA for cbb9f08

• Add safeCoalesce() to tests

  

  adamlerer

  authored and

  soumith

  committed
  f75ab85

  Copy full SHA for f75ab85

• Make coalesce() out of place

  

  adamlerer

  authored and

  soumith

  committed
  f290333

  Copy full SHA for f290333

• speed up accumulation

  

  adamlerer

  authored and

  soumith

  committed
  9643be7

  Copy full SHA for 9643be7

• Rename sparse tensor contiguous() to coalesce()

  

  adamlerer

  authored and

  soumith

  committed
  4f09461

  Copy full SHA for 4f09461

• Implement sparse pow. (#1387)

  

  ezyang

  authored and

  apaszke

  committed
  bafb2e5

  Copy full SHA for bafb2e5

• Documentation fix for torch.gather

  

  bunelr

  authored and

  apaszke

  committed
  28a7fbb

  Copy full SHA for 28a7fbb

• Refactor Python string utility function

  

  colesbury

  authored and

  apaszke

  committed
  4c1cdb6

  Copy full SHA for 4c1cdb6

• re-enable dilated convolutions on Kepler (#1394)

  

  ngimel

  authored and

  soumith

  committed
  775481e

  Copy full SHA for 775481e

• Merge commit '224f5eabf5cfb3a19abc1819f7dac230500b6bdb'

  soumith

  committed
  5b2aac7

  Copy full SHA for 5b2aac7

• half<->float conversion cleanup (#680)

  

  borisfom

  authored and

  soumith

  committed
  224f5ea

  Copy full SHA for 224f5ea

• Merge commit 'd6a31c68a0f39656257322a55c9e04dd579de828'

  soumith

  committed
  fd490c6

  Copy full SHA for fd490c6

• Add option to disable ppc64le's VSX support

  Show description for d6a31c6
  

  Bruno Rosa

  authored and

  soumith

  committed
  d6a31c6

  Copy full SHA for d6a31c6

• Add one more missing self.dilation parameter. (#1392)

  Show description for 96a281d
  

  ezyang

  authored and

  soumith

  committed
  96a281d

  Copy full SHA for 96a281d

• Allows dicts batches in dataloader. (#1354)

  Show description for 94b147f
  

  chsasank

  authored and

  soumith

  committed
  94b147f

  Copy full SHA for 94b147f

• guard topk for half (#759)

  

  killeent

  authored and

  soumith

  committed
  c26f687

  Copy full SHA for c26f687

• function -> lambda in test

  soumith

  committed
  8908000

  Copy full SHA for 8908000

• fix minor docs

  soumith

  committed
  8b1d572

  Copy full SHA for 8b1d572

• Add nn.Bilinear and tests

  

  uridah

  authored and

  soumith

  committed
  75f1989

  Copy full SHA for 75f1989

• Merge commit 'a44317fea88adddded91e068088415de1e66fd4b'

  soumith

  committed
  e221536

  Copy full SHA for e221536

• Change magma_sgesvd to magma_sgesdd which is significantly faster

  

  aam-at

  authored and

  soumith

  committed
  a44317f

  Copy full SHA for a44317f

• Revert "Parallelize TensorMethods.cpp builds (#1364)" (#1390)

  Show description for 24e5a90
  soumith

  authored
  24e5a90

  Copy full SHA for 24e5a90

• Parallelize TensorMethods.cpp builds (#1364)

  

  apaszke

  authored and

  soumith

  committed
  060048b

  Copy full SHA for 060048b

• make topk test unique

  soumith

  committed
  77035d1

  Copy full SHA for 77035d1

• enable topk for all cuda

  soumith

  committed
  50c9c23

  Copy full SHA for 50c9c23

• Merge commit '69574a6dc4036b0113c512a1b2d74e23682c8a3b'

  soumith

  committed
  3f81803

  Copy full SHA for 3f81803

• Merge commit '928f6516c16ff91c0a789d0a653551041d1bafd0'

  soumith

  committed
  d421c47

  Copy full SHA for d421c47

• implement expand/expandAs in CPU/GPU code

  

  killeent

  authored and

  soumith

  committed
  48f9e52

  Copy full SHA for 48f9e52

• implement expand/expandAs in CPU/GPU code

  

  killeent

  authored and

  soumith

  committed
  69574a6

  Copy full SHA for 69574a6

• implement expand/expandAs in CPU/GPU code

  

  killeent

  authored and

  soumith

  committed
  928f651

  Copy full SHA for 928f651

• Enable specifying of margin in HingeEmbeddingLoss (#1378)

  Show description for b93b525
  

  macaodha

  authored and

  soumith

  committed
  b93b525

  Copy full SHA for b93b525

• temp fix for transposed dilated convolution (#1388)

  

  ngimel

  authored and

  soumith

  committed
  8db2cf6

  Copy full SHA for 8db2cf6

Pagination

Previous

Next