Commits

Branch selector

 main

User selector

All users

All time

Commit History

Commits on Mar 19, 2026

• Revert "Add SymInt and SymFloat to Number (#177445)"

  Show description for 5f365c8
  pytorchmergebot

  committed
  5f365c8

  Copy full SHA for 5f365c8

• Revert "Fix type annotation for _sym_get_coordinate (#177446)"

  Show description for be7b8da
  pytorchmergebot

  committed
  be7b8da

  Copy full SHA for be7b8da

• [test][nv universal gemm][nvfp4] assert with `equal_nan=True` (#177475)

  Show description for 9aad0c0
  

  crcrpar

  authored and

  pytorchmergebot

  committed
  9aad0c0

  Copy full SHA for 9aad0c0

• [dynamo] Track TorchScriptObjectVariable as proxyable in HOPs and AC (#177785)

  Show description for 07e500b
  

  anijain2305

  authored and

  pytorchmergebot

  committed
  07e500b

  Copy full SHA for 07e500b

• Revert "[vllm hash update] update the pinned vllm hash (#177707)"

  Show description for 9594ae4
  pytorchmergebot

  committed
  9594ae4

  Copy full SHA for 9594ae4

• Revert "[triton hash update] update the pinned triton hash (#148492)"

  Show description for 5ec57c2
  pytorchmergebot

  committed
  5ec57c2

  Copy full SHA for 5ec57c2

• [BE][Ez]: Fuse empty strided_as call to empty_strided op (#177755)

  Show description for fefe931
  

  Skylion007

  authored and

  pytorchmergebot

  committed
  fefe931

  Copy full SHA for fefe931

• Revert "[user-streams] Fix check that didn't include torch.Event subclasses (#177701)"

  Show description for 08bf249
  pytorchmergebot

  committed
  08bf249

  Copy full SHA for 08bf249

• Fix type annotation for _sym_get_coordinate (#177446)

  Show description for b8d53c6
  

  aorenste

  authored and

  pytorchmergebot

  committed
  b8d53c6

  Copy full SHA for b8d53c6

• Update kineto submodule to 0c52fa62fe31 (#177753)

  Show description for 42244bc
  

  scotts

  authored and

  pytorchmergebot

  committed
  42244bc

  Copy full SHA for 42244bc

• [2/2] Introduce torch.accelerator.Graph as a unified frontend Graph interface (#171285)

  Show description for 68050bc
  

  guangyey

  authored and

  pytorchmergebot

  committed
  68050bc

  Copy full SHA for 68050bc

• [triton hash update] update the pinned triton hash (#148492)

  Show description for 5de4dd3
  

  pytorchupdatebot

  authored and

  pytorchmergebot

  committed
  5de4dd3

  Copy full SHA for 5de4dd3

• [torchtitan hash update] update the pinned torchtitan hash (#177706)

  Show description for b583eff
  

  pytorchupdatebot

  authored and

  pytorchmergebot

  committed
  b583eff

  Copy full SHA for b583eff

• [vision hash update] update the pinned vision hash (#175593)

  Show description for 9c62c48
  

  pytorchupdatebot

  authored and

  pytorchmergebot

  committed
  9c62c48

  Copy full SHA for 9c62c48

• [vllm hash update] update the pinned vllm hash (#177707)

  Show description for 1987efa
  

  pytorchupdatebot

  authored and

  pytorchmergebot

  committed
  1987efa

  Copy full SHA for 1987efa

• [inductor] run small dce after pattern match so it does not insert dead code (#177547)

  Show description for f394549
  

  v0i0

  authored and

  pytorchmergebot

  committed
  f394549

  Copy full SHA for f394549

• [inductor-pallas] Fix TPU transformer layer test tolerances (#177752)

  Show description for 55b2a52
  

  v0i0

  authored and

  pytorchmergebot

  committed
  55b2a52

  Copy full SHA for 55b2a52

• [coor-targets] Accept ProcessGroup directly in functional collective op schemas (#172795)

  Show description for 12bfd2c
  

  aorenste

  authored and

  pytorchmergebot

  committed
  12bfd2c

  Copy full SHA for 12bfd2c

• [ROCm] Enable hipSPARSELt for ROCm >= 7.12 (#170852)

  Show description for 0385f89
  

  rraminen

  authored and

  pytorchmergebot

  committed
  0385f89

  Copy full SHA for 0385f89

Commits on Mar 18, 2026

• Add pre_grad_pass_timing config for early vs late pre-grad passes (#177429)

  Show description for 2a90511
  

  frgossen

  authored and

  pytorchmergebot

  committed
  2a90511

  Copy full SHA for 2a90511

• Remove unused static_inputs_log from aot_autograd.py (#177428)

  Show description for 8f38cad
  

  frgossen

  authored and

  pytorchmergebot

  committed
  8f38cad

  Copy full SHA for 8f38cad

• Add UUID-based cache key support for pre-grad custom passes (#177403)

  Show description for 00fa5de
  

  frgossen

  authored and

  pytorchmergebot

  committed
  00fa5de

  Copy full SHA for 00fa5de

• Add AArch64 xfails for inductor, nn, jit, and linalg tests (#177584)

  Show description for ef99230
  

  robert-hardwick

  authored and

  pytorchmergebot

  committed
  ef99230

  Copy full SHA for ef99230

• Revert "[dynamo] Deprecate inline_inbuilt_nn_modules config (#177490)"

  Show description for 5091ccd
  pytorchmergebot

  committed
  5091ccd

  Copy full SHA for 5091ccd

• [dynamo] Fix torch.Size dict lookups with tensor-backed keys (#177313)

  Show description for 5a4d055
  

  bobrenjc93

  authored and

  pytorchmergebot

  committed
  5a4d055

  Copy full SHA for 5a4d055

• Revert "[xpu][test][1/N] Enable tests of test_nn.py on Intel GPU - instantiate TestNN with instantiate_device_type_tests (#166396)"

  Show description for 4c32aea
  pytorchmergebot

  committed
  4c32aea

  Copy full SHA for 4c32aea

• `torch.bcomplex32` shell dtype. (#173783)

  Show description for 778cf35
  

  hameerabbasi

  authored and

  pytorchmergebot

  committed
  778cf35

  Copy full SHA for 778cf35

• [dynamo] Deprecate inline_inbuilt_nn_modules config (#177490)

  Show description for 95ef471
  

  anijain2305

  authored and

  pytorchmergebot

  committed
  95ef471

  Copy full SHA for 95ef471

• Fix behavor of CTC loss cuDNN backend when `zero_infinity=True` (#176911)

  Show description for 0f1438d
  

  bencwallace

  authored and

  pytorchmergebot

  committed
  0f1438d

  Copy full SHA for 0f1438d

• Add cross-process AOT autograd cache hit test (#177397)

  Show description for f083dc2
  

  frgossen

  authored and

  pytorchmergebot

  committed
  f083dc2

  Copy full SHA for f083dc2

• [user-streams] Fix check that didn't include torch.Event subclasses (#177701)

  Show description for a4a0059
  

  mlazos

  authored and

  pytorchmergebot

  committed
  a4a0059

  Copy full SHA for a4a0059

• [MPS] Migrate minimum/maximum from MPSGraph to native Metal (#177747)

  Show description for 692526f
  

  2 peopleauthored and

  pytorchmergebot

  committed
  692526f

  Copy full SHA for 692526f

• [functorch] Fix grad/vjp/jvp returning zeros under inference_mode (#177596)

  Show description for 05710a7
  

  stmcgovern

  authored and

  pytorchmergebot

  committed
  05710a7

  Copy full SHA for 05710a7

• [memoryviz] Add a tab to show private pool memory view (#177289)

  Show description for 7a31ab9
  

  yushangdi

  authored and

  pytorchmergebot

  committed
  7a31ab9

  Copy full SHA for 7a31ab9

• Add pool_id to memory snapshot trace entries (#177717)

  Show description for 079faf3
  

  ngimel

  authored and

  pytorchmergebot

  committed
  079faf3

  Copy full SHA for 079faf3