Commits

Breadcrumbs

History for

onmain

User selector

All users

Datepicker

All time

Commit History

Commits on Apr 26, 2024

• Avoid COW materialize in various places (1) (pytorch#124984)

  Show description for abcb42c
  

  kurtamohler

  authored and

  pytorchmergebot

  committedApr 26, 2024
  abcb42c

  Copy full SHA for abcb42c

• [Aten] Fix XPU convolution_overrideable input memory format. (pytorch#124841)

  Show description for 36af9c0
  

  etaf

  authored and

  pytorchmergebot

  committedApr 26, 2024
  36af9c0

  Copy full SHA for 36af9c0

Commits on Apr 25, 2024

• [MPS][BE] Error-check linear (pytorch#124952)

  Show description for db3a2d7
  

  malfet

  authored and

  pytorchmergebot

  committedApr 25, 2024
  db3a2d7

  Copy full SHA for db3a2d7

• [CUDA] Fix 64-bit indexing in vol2col in conv3d (pytorch#124650)

  Show description for 973d724
  

  eqy

  authored and

  pytorchmergebot

  committedApr 25, 2024
  973d724

  Copy full SHA for 973d724

• Revert "[CUDA] Fix 64-bit indexing in vol2col in conv3d (pytorch#124650)"

  Show description for 24ed909
  pytorchmergebot

  committedApr 25, 2024
  24ed909

  Copy full SHA for 24ed909

• Revert "[cuDNN] cuDNN SDPA (Flash Attention) Backward (pytorch#122510)"

  Show description for 7a6813b
  pytorchmergebot

  committedApr 25, 2024
  7a6813b

  Copy full SHA for 7a6813b

• [foreach] Use same dtypes when dtypesIfCUDA is None (pytorch#124813)

  Show description for b21bf5e
  

  crcrpar

  authored and

  pytorchmergebot

  committedApr 25, 2024
  b21bf5e

  Copy full SHA for b21bf5e

Commits on Apr 24, 2024

• [ROCm][Inductor] Disable conv cache emptying with hipgraphs (pytorch#124791)

  Show description for a248c24
  

  pragupta

  authored and

  pytorchmergebot

  committedApr 24, 2024
  a248c24

  Copy full SHA for a248c24

• [CUDA] Fix 64-bit indexing in vol2col in conv3d (pytorch#124650)

  Show description for 71d92ba
  

  eqy

  authored and

  pytorchmergebot

  committedApr 24, 2024
  71d92ba

  Copy full SHA for 71d92ba

• [ROCm] Fix Int_mm() Integration with hipblasLT (pytorch#122431)

  Show description for 2e7b8ff
  

  petrex

  authored and

  pytorchmergebot

  committedApr 24, 2024
  2e7b8ff

  Copy full SHA for 2e7b8ff

• Add Half support to torch.sparse.addmm for CPU (pytorch#124694)

  Show description for 57a12d9
  

  yanbing-j

  authored and

  pytorchmergebot

  committedApr 24, 2024
  57a12d9

  Copy full SHA for 57a12d9

• [ROCm] avoid heap buffer overflow in hiprtc failure logs (pytorch#121865)

  Show description for 1ab0b3c
  

  jeffdaily

  authored and

  pytorchmergebot

  committedApr 24, 2024
  1ab0b3c

  Copy full SHA for 1ab0b3c

• [cuDNN] cuDNN SDPA (Flash Attention) Backward (pytorch#122510)

  Show description for 64af899
  

  eqy

  authored and

  pytorchmergebot

  committedApr 24, 2024
  64af899

  Copy full SHA for 64af899

Commits on Apr 23, 2024

• [Dynamo] Add dynamo support to torch.func.linearize (pytorch#123118)

  Show description for 763dc26
  

  guilhermeleobas

  authored and

  pytorchmergebot

  committedApr 23, 2024
  763dc26

  Copy full SHA for 763dc26

• add fused_sgd_kernel support for CPU device (pytorch#123629)

  Show description for 3c964ad
  

  zhuhaozhe

  authored and

  pytorchmergebot

  committedApr 23, 2024
  3c964ad

  Copy full SHA for 3c964ad

Commits on Apr 22, 2024

• Intel GPU oneDNN Upstreaming: Convolution operators support (pytorch#117529)

  Show description for ad7b5d3
  

  ZhiweiYan-96

  authored and

  pytorchmergebot

  committedApr 22, 2024
  ad7b5d3

  Copy full SHA for ad7b5d3

• Intel GPU oneDNN upstreaming: Conv primitive integration (pytorch#117512)

  Show description for 9d4bef6
  

  ZhiweiYan-96

  authored and

  pytorchmergebot

  committedApr 22, 2024
  9d4bef6

  Copy full SHA for 9d4bef6

Commits on Apr 21, 2024

• [MPS] Fixes GELU, LeakyRELU and MISH on non-contiguous tensors (pytorch#123049)

  Show description for a6a3f2e
  

  jtang98

  authored and

  pytorchmergebot

  committedApr 21, 2024
  a6a3f2e

  Copy full SHA for a6a3f2e

Commits on Apr 19, 2024

• Prep PR for cutlass 3.5 update (pytorch#124412)

  Show description for d8a98dd
  

  drisspg

  authored and

  pytorchmergebot

  committedApr 19, 2024
  d8a98dd

  Copy full SHA for d8a98dd

• Int4MM: Unswizzle for different dtypes (pytorch#124448)

  Show description for c74dfca
  

  malfet

  authored and

  pytorchmergebot

  committedApr 19, 2024
  c74dfca

  Copy full SHA for c74dfca

• Fix compilation on aarch64 with gcc (pytorch#124511)

  Show description for e6a788a
  

  malfet

  authored and

  pytorchmergebot

  committedApr 19, 2024
  e6a788a

  Copy full SHA for e6a788a

• [AMD] TunableOp take priority over DISABLE_ADDMM_HIP_LT (pytorch#124161)

  Show description for 661fd23
  

  xw285cornell

  authored and

  pytorchmergebot

  committedApr 19, 2024
  661fd23

  Copy full SHA for 661fd23

• [AMD] Remove deprecated macro from COnvUtils (pytorch#124158)

  Show description for 8869b54
  

  xw285cornell

  authored and

  pytorchmergebot

  committedApr 19, 2024
  8869b54

  Copy full SHA for 8869b54

• [sparse] Add fast semi-structured spasification kernels (pytorch#122350)

  Show description for c9db59e
  

  jcaip

  authored and

  pytorchmergebot

  committedApr 19, 2024
  c9db59e

  Copy full SHA for c9db59e

• Add vectorized norm fill for ppc64le (pytorch#113351)

  Show description for 88fa843
  

  pratiklp00

  authored and

  pytorchmergebot

  committedApr 19, 2024
  88fa843

  Copy full SHA for 88fa843

• [optim] add fused_adam/adamw_kernel support for CPU device (pytorch#123074)

  Show description for b412b75
  

  janeyx99

  authored and

  pytorchmergebot

  committedApr 19, 2024
  b412b75

  Copy full SHA for b412b75

• Fix AVX2 int4pack_mm_kernel crash if weighs are unaligned (pytorch#124433)

  Show description for b2f6cfd
  

  malfet

  authored and

  pytorchmergebot

  committedApr 19, 2024
  b2f6cfd

  Copy full SHA for b2f6cfd

Commits on Apr 18, 2024

• Make copy_cast, softmax and cat_out unranked (pytorch#123191)

  Show description for e0792cf
  

  kulinseth

  authored and

  pytorchmergebot

  committedApr 18, 2024
  e0792cf

  Copy full SHA for e0792cf

• Fixed issue in affine_grid_backward when grad_grid is non-contiguous (pytorch#124370)

  Show description for 415a8f6
  

  vfdev-5

  authored and

  pytorchmergebot

  committedApr 18, 2024
  415a8f6

  Copy full SHA for 415a8f6

• [MPS] Fix crash with binary_cross_entropy is invoked for half dtypes (pytorch#124258)

  Show description for 5677128
  

  malfet

  authored and

  pytorchmergebot

  committedApr 18, 2024
  5677128

  Copy full SHA for 5677128

• [ATen] Add CPU fp16 support for nll_loss and cross_entropy_loss (pytorch#123256)

  Show description for 6fcbeb3
  

  etaf

  authored and

  pytorchmergebot

  committedApr 18, 2024
  6fcbeb3

  Copy full SHA for 6fcbeb3

Commits on Apr 17, 2024

• Extend int[48]mm ops to float32 input (pytorch#124287)

  Show description for 00372b1
  

  malfet

  authored and

  pytorchmergebot

  committedApr 17, 2024
  00372b1

  Copy full SHA for 00372b1

• [Intel GPU] oneDNN GPU GEMM support (pytorch#117202)

  Show description for 9875a83
  

  ZhiweiYan-96

  authored and

  pytorchmergebot

  committedApr 17, 2024
  9875a83

  Copy full SHA for 9875a83

• Intel GPU oneDNN upstreaming for primitive integration (pytorch#117112)

  Show description for cc18afa
  

  ZhiweiYan-96

  authored and

  pytorchmergebot

  committedApr 17, 2024
  cc18afa

  Copy full SHA for cc18afa

• Speedup int4mm_kernel with NEON (pytorch#124257)

  Show description for 46324fe
  

  malfet

  authored and

  pytorchmergebot

  committedApr 17, 2024
  46324fe

  Copy full SHA for 46324fe

Pagination

Previous

Next