Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Sep 19, 2025

• Rm pytorch deps platform args (#163086)

  Show description for 5f25dbe
  

  jaejunku

  authored and

  pytorchmergebot

  committed
  5f25dbe

  Copy full SHA for 5f25dbe

• Update torch-xpu-ops commit pin (#163244)

  Show description for e134bb3
  

  Chao1Han

  authored and

  pytorchmergebot

  committed
  e134bb3

  Copy full SHA for e134bb3

• add more restriction to fusion with large accumulate reads (#163163)

  Show description for 6e680ae
  

  xuanzhang816

  authored and

  pytorchmergebot

  committed
  6e680ae

  Copy full SHA for 6e680ae

• Refactor PrecompileContext to be considerably more debuggable (#162740)

  Show description for 3c9e220
  

  jamesjwu

  authored and

  pytorchmergebot

  committed
  3c9e220

  Copy full SHA for 3c9e220

• [ROCm] Bump FBGEMM commit to avoid CK errors (#162590)

  Show description for c9b80c4
  

  pragupta

  authored and

  pytorchmergebot

  committed
  c9b80c4

  Copy full SHA for c9b80c4

• [CP][BE] Correct the names of some tests (#162541)

  Show description for cd4303a
  

  fegin

  authored and

  pytorchmergebot

  committed
  cd4303a

  Copy full SHA for cd4303a

• [CI] Move Windows build/tests to Python-3.10 (#162862)

  Show description for 2dcd153
  

  malfet

  authored and

  pytorchmergebot

  committed
  2dcd153

  Copy full SHA for 2dcd153

• Change DebugMode record_torchfunction default to False (#163293)

  Show description for 04842ac
  

  SherlockNoMad

  authored and

  pytorchmergebot

  committed
  04842ac

  Copy full SHA for 04842ac

Commits on Sep 18, 2025

• Deprecate Lite Interpreter (#163289)

  Show description for 17c1653
  

  JacobSzwejbka

  authored and

  pytorchmergebot

  committed
  17c1653

  Copy full SHA for 17c1653

• [functional] Use the saved device on storage instead for device_custom (#162987)

  Show description for ddc56f6
  

  anijain2305

  authored and

  pytorchmergebot

  committed
  ddc56f6

  Copy full SHA for ddc56f6

• [CP] Remove the need of recording cp_dim in the global var (#162540)

  Show description for 096d35c
  

  fegin

  authored and

  pytorchmergebot

  committed
  096d35c

  Copy full SHA for 096d35c

• [dynamic shapes] DynamicInts prototype (#162194)

  Show description for 4c00707
  

  pianpwk

  authored and

  pytorchmergebot

  committed
  4c00707

  Copy full SHA for 4c00707

• Try updating ET pin in PT/PT (#159664)

  Show description for f4eca0e
  

  mergennachin

  authored and

  pytorchmergebot

  committed
  f4eca0e

  Copy full SHA for f4eca0e

• Turn on capture_dynamic_output_shape_ops when fullgraph=True (#163123)

  Show description for ed3438f
  

  bobrenjc93

  authored and

  pytorchmergebot

  committed
  ed3438f

  Copy full SHA for ed3438f

• Turn on capture_scalar_outputs when fullgraph=True (#163121)

  Show description for 7dcb568
  

  bobrenjc93

  authored and

  pytorchmergebot

  committed
  7dcb568

  Copy full SHA for 7dcb568

• [DTensor] Fix DTensor.mean with uneven sharding (#163241)

  Show description for bb7c9a2
  

  SherlockNoMad

  authored and

  pytorchmergebot

  committed
  bb7c9a2

  Copy full SHA for bb7c9a2

• test: ensure editable cached wrapper is respected (#160943)

  Show description for 159c214
  

  ezyang

  authored and

  pytorchmergebot

  committed
  159c214

  Copy full SHA for 159c214

• [ROCm] update ci_expected_accuracy for dynamo benchmarks (#163256)

  Show description for 62a746f
  

  jeffdaily

  authored and

  pytorchmergebot

  committed
  62a746f

  Copy full SHA for 62a746f

• Add local file path to inductor_output_code trace metadata (#160920)

  Show description for 8627454
  

  ezyang

  authored and

  pytorchmergebot

  committed
  8627454

  Copy full SHA for 8627454

• [unit test] correct wrong input shape in test_flop_fx (#163148)

  Show description for 93964ed
  

  thenumberouscode

  authored and

  pytorchmergebot

  committed
  93964ed

  Copy full SHA for 93964ed

• [SymmMem] Fix put_signal + wait_until hang (#163194)

  Show description for 80f8be9
  

  kwen2501

  authored and

  pytorchmergebot

  committed
  80f8be9

  Copy full SHA for 80f8be9

• Massive hack to make autograd shut up about threaded PG mutations (#163238)

  Show description for e36a6fc
  

  ezyang

  authored and

  pytorchmergebot

  committed
  e36a6fc

  Copy full SHA for e36a6fc

• [WOQ] Integrate CUDA support for concat linear int8pack_mm woq optimization pattern (#161848)

  Show description for 23af32a
  

  bbeckca

  authored and

  pytorchmergebot

  committed
  23af32a

  Copy full SHA for 23af32a

• [submodule] CUTLASS upgrade to 4.2.0 and change cutlass to cutlass_cppgen (#163092)

  Show description for a81a2e5
  

  henrylhtsang

  authored and

  pytorchmergebot

  committed
  a81a2e5

  Copy full SHA for a81a2e5

• Revert "[torch][cuda][device_limits] Library for querying device hardware limits for flops and bandwidth (#162942)"

  Show description for 4b7aed8
  pytorchmergebot

  committed
  4b7aed8

  Copy full SHA for 4b7aed8

• Move prioritized text linker optimization code from setup.py to cmake (#160078)

  Show description for 1aeac30
  

  robert-hardwick

  authored and

  pytorchmergebot

  committed
  1aeac30

  Copy full SHA for 1aeac30

• Don't register wrong overload to prim decomp (#163138)

  Show description for 56893ca
  

  tugsbayasgalan

  authored and

  pytorchmergebot

  committed
  56893ca

  Copy full SHA for 56893ca

• [CI] reuse old whl: fix metadata file not getting version replaced (#163214)

  Show description for af8c232
  

  clee2000

  authored and

  pytorchmergebot

  committed
  af8c232

  Copy full SHA for af8c232

• [testing] Add test owner labels for some ao sparse tests (#163203)

  Show description for 4908fb5
  

  clee2000

  authored and

  pytorchmergebot

  committed
  4908fb5

  Copy full SHA for 4908fb5

• support unbacked softmax / logsoftmax (#162216)

  Show description for 3c8b905
  

  ColinPeppler

  authored and

  pytorchmergebot

  committed
  3c8b905

  Copy full SHA for 3c8b905

• fixing graph break for namedtuple._replace (#160139)

  Show description for 1f21f85
  

  morrison-turnansky

  authored and

  pytorchmergebot

  committed
  1f21f85

  Copy full SHA for 1f21f85

• Update Microsoft C++ Redistributable to the latest version (#161430)

  Show description for 1330c63
  

  saman-amd

  authored and

  pytorchmergebot

  committed
  1330c63

  Copy full SHA for 1330c63

• [ROCm] test_aot_inductor: Enable fp8 tests. (#163050)

  Show description for 8bc4a46
  

  jagadish-amd

  authored and

  pytorchmergebot

  committed
  8bc4a46

  Copy full SHA for 8bc4a46

• [ROCm] Remove HIPBLASLT_ALLOW_TF32 from codebase (#162998)

  Show description for e769026
  

  2 peopleauthored and

  pytorchmergebot

  committed
  e769026

  Copy full SHA for e769026

• Reland #161649, vectorize stored in cat for all dtypes (#162440)

  Show description for 14f8d86
  

  ngimel

  authored and

  pytorchmergebot

  committed
  14f8d86

  Copy full SHA for 14f8d86

Pagination

Previous

Next