Commits

Branch selector

 master

User selector

gchanan

Datepicker

All time

Commit History

Commits on May 13, 2021

• Fix legacy tensor constructor/new matching incorrect signature with d… (pytorch#58108)

  Show description for e8574b8
  

  gchanan

  authored and

  facebook-github-bot

  committed
  e8574b8

  Copy full SHA for e8574b8

Commits on Feb 18, 2021

• Allow broadcasting against lerp weights. (pytorch#52319)

  Show description for 983347f
  

  gchanan

  authored and

  facebook-github-bot

  committed
  983347f

  Copy full SHA for 983347f

• Fix upsample bicubic2d batching handling on CPU. (pytorch#52389)

  Show description for f72b4b8
  

  gchanan

  authored and

  facebook-github-bot

  committed
  f72b4b8

  Copy full SHA for f72b4b8

Commits on Jan 12, 2021

• Stop moving scalars to GPU for one computation in leaky_rrelu_backward. (pytorch#50115)

  Show description for c198e6c
  

  gchanan

  authored and

  facebook-github-bot

  committed
  c198e6c

  Copy full SHA for c198e6c

• Stop using an unnecessary scalar_to_tensor(..., device) call. (pytorch#50114)

  Show description for b001c4c
  

  gchanan

  authored and

  facebook-github-bot

  committed
  b001c4c

  Copy full SHA for b001c4c

Commits on Jan 11, 2021

• Move scalar_to_tensor_default_dtype out of ScalarOps.h because it's only useful for torch.where. (pytorch#50111)

  Show description for 0f412aa
  

  gchanan

  authored and

  facebook-github-bot

  committed
  0f412aa

  Copy full SHA for 0f412aa

Commits on Jan 8, 2021

• Stop using c10::scalar_to_tensor in float_power. (pytorch#50105)

  Show description for 88bd69b
  

  gchanan

  authored and

  facebook-github-bot

  committed
  88bd69b

  Copy full SHA for 88bd69b

Commits on Jan 4, 2021

• torch.xlogy: Use wrapped_scalar_tensor / gpu_with_scalars to speed up GPU kernel. (pytorch#49926)

  Show description for 74dcb6d
  

  gchanan

  authored and

  facebook-github-bot

  committed
  74dcb6d

  Copy full SHA for 74dcb6d

Commits on Dec 29, 2020

• Update gather documentation to allow index.shape[k] <= input.shape[k] rather than ==. (pytorch#41887)

  Show description for d434ac3
  

  gchanan

  authored and

  facebook-github-bot

  committed
  d434ac3

  Copy full SHA for d434ac3

Commits on Nov 12, 2020

• Revert "Fixed einsum compatibility/performance issues (pytorch#46398)" (pytorch#47821)

  Show description for b6cb2ca
  

  gchanan

  authored and

  facebook-github-bot

  committed
  b6cb2ca

  Copy full SHA for b6cb2ca

Commits on Nov 10, 2020

• Fix type promotion for trace on CPU. (pytorch#47305)

  Show description for 65a72ca
  

  gchanan

  authored and

  facebook-github-bot

  committed
  65a72ca

  Copy full SHA for 65a72ca

Commits on Nov 5, 2020

• __noinline__ the top level igamma cuda kernel. (pytorch#47414)

  Show description for 1519c71
  

  gchanan

  authored and

  facebook-github-bot

  committed
  1519c71

  Copy full SHA for 1519c71

• Move igamma cuda specific code to kernel file. (pytorch#47410)

  Show description for ae374dc
  

  gchanan

  authored and

  facebook-github-bot

  committed
  ae374dc

  Copy full SHA for ae374dc

• Split IGamma cuda kernel into it's own file to speed up compilation times. (pytorch#47401)

  Show description for 32c76db
  

  gchanan

  authored and

  facebook-github-bot

  committed
  32c76db

  Copy full SHA for 32c76db

• Clean up some imports in cuda kernel code. (pytorch#47400)

  Show description for ff3e1de
  

  gchanan

  authored and

  facebook-github-bot

  committed
  ff3e1de

  Copy full SHA for ff3e1de

Commits on Nov 4, 2020

• Split up BinaryMiscOpKernels.cu because it's slow to compile. (pytorch#47362)

  Show description for da491d7
  

  gchanan

  authored and

  facebook-github-bot

  committed
  da491d7

  Copy full SHA for da491d7

Commits on Oct 15, 2020

• Stop running clang-tidy on torch/csrc/generic/*.cpp. (pytorch#46335)

  Show description for 7580962
  

  gchanan

  authored and

  facebook-github-bot

  committed
  7580962

  Copy full SHA for 7580962

Commits on Oct 12, 2020

• Improve error checking of Storage._writeFile. (pytorch#46036)

  Show description for 2070834
  

  gchanan

  authored and

  facebook-github-bot

  committed
  2070834

  Copy full SHA for 2070834

Commits on Sep 28, 2020

• Remove CriterionTest.test_cuda code for dtype None. (pytorch#45316)

  Show description for 1097fe0
  

  gchanan

  authored and

  facebook-github-bot

  committed
  1097fe0

  Copy full SHA for 1097fe0

• Remove convert_target from NN tests. (pytorch#45291)

  Show description for 5d1fee2
  

  gchanan

  authored and

  facebook-github-bot

  committed
  5d1fee2

  Copy full SHA for 5d1fee2

Commits on Sep 21, 2020

• Turn on gradgrad check for BCELoss Criterion Tests. (pytorch#44894)

  Show description for a6895d4
  

  gchanan

  authored and

  facebook-github-bot

  committed
  a6895d4

  Copy full SHA for a6895d4

Commits on Sep 18, 2020

• Stop using check_criterion_jacobian. (pytorch#44786)

  Show description for 07b7e44
  

  gchanan

  authored and

  facebook-github-bot

  committed
  07b7e44

  Copy full SHA for 07b7e44

• Stop ignoring errors in cuda nn module tests. (pytorch#44783)

  Show description for 6d178f6
  

  gchanan

  authored and

  facebook-github-bot

  committed
  6d178f6

  Copy full SHA for 6d178f6

Commits on Sep 17, 2020

• Always use NewModuleTest instead of ModuleTest. (pytorch#44745)

  Show description for 503c748
  

  gchanan

  authored and

  facebook-github-bot

  committed
  503c748

  Copy full SHA for 503c748

Commits on Sep 11, 2020

• Simplify target handling in nn gradcheck. (pytorch#44507)

  Show description for 192c411
  

  gchanan

  authored and

  facebook-github-bot

  committed
  192c411

  Copy full SHA for 192c411

• Fix SmoothL1Loss when target.requires_grad is True. (pytorch#44486)

  Show description for 5579b53
  

  gchanan

  authored and

  facebook-github-bot

  committed
  5579b53

  Copy full SHA for 5579b53

• Fix L1Loss when target.requires_grad is True. (pytorch#44471)

  Show description for 3de2c0b
  

  gchanan

  authored and

  facebook-github-bot

  committed
  3de2c0b

  Copy full SHA for 3de2c0b

• Fix MSELoss when target.requires_grad is True. (pytorch#44437)

  Show description for d07d25a
  

  gchanan

  authored and

  facebook-github-bot

  committed
  d07d25a

  Copy full SHA for d07d25a

Commits on Sep 10, 2020

• Merge criterion_tests and new_criterion_tests. (pytorch#44398)

  Show description for c8914af
  

  gchanan

  authored and

  facebook-github-bot

  committed
  c8914af

  Copy full SHA for c8914af

• Combine criterion and new criterion tests in test_jit. (pytorch#43958)

  Show description for fa158c4
  

  gchanan

  authored and

  facebook-github-bot

  committed
  fa158c4

  Copy full SHA for fa158c4

• Stop ignoring NotImplementedErrors in cuda CriterionTests. (pytorch#44381)

  Show description for af9cad7
  

  gchanan

  authored and

  facebook-github-bot

  committed
  af9cad7

  Copy full SHA for af9cad7

Commits on Sep 3, 2020

• For CriterionTests, have check_gradgrad actually only affect gradgrad checks. (pytorch#44060)

  Show description for 49215d7
  

  gchanan

  authored and

  facebook-github-bot

  committed
  49215d7

  Copy full SHA for 49215d7

• Rename NewCriterionTest to CriterionTest. (pytorch#44056)

  Show description for 5973b44
  

  gchanan

  authored and

  facebook-github-bot

  committed
  5973b44

  Copy full SHA for 5973b44

• Merge CriterionTest into NewCriterionTest. (pytorch#44055)

  Show description for cae52b4
  

  gchanan

  authored and

  facebook-github-bot

  committed
  cae52b4

  Copy full SHA for cae52b4

• Allow criterion backwards test on modules requiring extra args (i.e. CTCLoss). (pytorch#44050)

  Show description for 68a1fbe
  

  gchanan

  authored and

  facebook-github-bot

  committed
  68a1fbe

  Copy full SHA for 68a1fbe

Pagination

Previous

Next