Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on May 2, 2025

• [MPS][BE] Do not dispatch empty kernels (#152663)

  Show description for cb0cf7e
  

  malfet

  authored and

  pytorchmergebot

  committed
  cb0cf7e

  Copy full SHA for cb0cf7e

• Revert "[cutlass backend] Minor lru_cache to slightly speed up filtering ops (#152577)"

  Show description for 50d4698
  pytorchmergebot

  committed
  50d4698

  Copy full SHA for 50d4698

• Enable -Wunused on torch targets (#150077)

  Show description for e9e1aac
  

  cyyever

  authored and

  pytorchmergebot

  committed
  e9e1aac

  Copy full SHA for e9e1aac

• Fix: Consider input defined unbacked during inductor codegen for runtime asserts (#152231)

  Show description for 38a9a8b
  

  laithsakka

  authored and

  pytorchmergebot

  committed
  38a9a8b

  Copy full SHA for 38a9a8b

• [SymmMem] Add all_to_all_vdev (#151819)

  Show description for 829752b
  

  kwen2501

  authored and

  pytorchmergebot

  committed
  829752b

  Copy full SHA for 829752b

• Revert "Enable -Wunused on torch targets (#150077)"

  Show description for 6dadfc4
  pytorchmergebot

  committed
  6dadfc4

  Copy full SHA for 6dadfc4

• [inductor][invoke_subgraph] Remove assertion checks for outputs of invoke_subgraph (#152384)

  Show description for 3731b70
  

  anijain2305

  authored and

  pytorchmergebot

  committed
  3731b70

  Copy full SHA for 3731b70

• [invoke_subgraph] rename identifiers to prevent python mangling (#152581)

  Show description for 9e3fc41
  

  anijain2305

  authored and

  pytorchmergebot

  committed
  9e3fc41

  Copy full SHA for 9e3fc41

• Revert "Use swap_tensors path in nn.Module.to for all subclasses that override __torch_dispatch__ (#152539)"

  Show description for 4f9f1ab
  pytorchmergebot

  committed
  4f9f1ab

  Copy full SHA for 4f9f1ab

• [SymmMem] Add all-to-all (#151498)

  Show description for d7961a1
  

  kwen2501

  authored and

  pytorchmergebot

  committed
  d7961a1

  Copy full SHA for d7961a1

• Revert "[Inductor] Add decomposeK as an autotuning choice for mm (#150654)"

  Show description for 7c3e679
  pytorchmergebot

  committed
  7c3e679

  Copy full SHA for 7c3e679

• [invoke_subgraph] Unpacked operands (#152547)

  Show description for 4649fd1
  

  anijain2305

  authored and

  pytorchmergebot

  committed
  4649fd1

  Copy full SHA for 4649fd1

• Revert "[BE] Update numba versions (#152557)"

  Show description for e6989ce
  pytorchmergebot

  committed
  e6989ce

  Copy full SHA for e6989ce

• Remove unnecessary __STDC_FORMAT_MACROS macro (#152513)

  Show description for ac5de6d
  

  fffrog

  authored and

  pytorchmergebot

  committed
  ac5de6d

  Copy full SHA for ac5de6d

• [CUDAGraph Trees] support memory allocation on side stream (#152472)

  Show description for d969e2e
  

  BoyuanFeng

  authored and

  pytorchmergebot

  committed
  d969e2e

  Copy full SHA for d969e2e

• [ez] fix grammar mistakes in StatefulSymbolicContext comment (#152598)

  Show description for 1f89865
  

  bobrenjc93

  authored and

  pytorchmergebot

  committed
  1f89865

  Copy full SHA for 1f89865

• [CP] Fix the offsets to KV in backward (#152625)

  Show description for 36e5ff6
  

  fegin

  authored and

  pytorchmergebot

  committed
  36e5ff6

  Copy full SHA for 36e5ff6

• [cutlass backend] Minor lru_cache to slightly speed up filtering ops (#152577)

  Show description for 1fef3cd
  

  henrylhtsang

  authored and

  pytorchmergebot

  committed
  1fef3cd

  Copy full SHA for 1fef3cd

• [refactor] refactor dense implementation of auto_functionalized_v2 for better clarity (#152248)

  Show description for 5b59389
  

  ydwu4

  authored and

  pytorchmergebot

  committed
  5b59389

  Copy full SHA for 5b59389

• [hop] make materialize_as_graph's include and exclude dispatch key set optional (#152247)

  Show description for 380327c
  

  ydwu4

  authored and

  pytorchmergebot

  committed
  380327c

  Copy full SHA for 380327c

• [hop][schema] allow adding kw_only info to schema argument (#152246)

  Show description for a776a56
  

  ydwu4

  authored and

  pytorchmergebot

  committed
  a776a56

  Copy full SHA for a776a56

• [hop][be] make check_input_alias_and_mutation_return_ouputs create new fake mode (#152245)

  Show description for 7e7b9ca
  

  ydwu4

  authored and

  pytorchmergebot

  committed
  7e7b9ca

  Copy full SHA for 7e7b9ca

• [BE] Update numba versions (#152557)

  Show description for b5995cb
  

  malfet

  authored and

  pytorchmergebot

  committed
  b5995cb

  Copy full SHA for b5995cb

• [Environment Variable][Rebase] Use thread-safe getenv functions (#140200)

  Show description for ce94b21
  

  cyyever

  authored and

  pytorchmergebot

  committed
  ce94b21

  Copy full SHA for ce94b21

• [ONNX] Delete JitTraceConvertStrategy (#152556)

  Show description for a5dd701
  

  titaiwangms

  authored and

  pytorchmergebot

  committed
  a5dd701

  Copy full SHA for a5dd701

• [inductor] if unbacked symint in old-size or new-size skip mark_reuse check (#152379)

  Show description for 3c54e0c
  

  ColinPeppler

  authored and

  pytorchmergebot

  committed
  3c54e0c

  Copy full SHA for 3c54e0c

Commits on May 1, 2025

• [Inductor] Add decomposeK as an autotuning choice for mm (#150654)

  Show description for fdcfc6a
  

  PaulZhang12

  authored and

  pytorchmergebot

  committed
  fdcfc6a

  Copy full SHA for fdcfc6a

• Fix some inductor periodic benchmarks (#152605)

  Show description for 64957db
  

  zou3519

  authored and

  pytorchmergebot

  committed
  64957db

  Copy full SHA for 64957db

• [dynamo][ca] support dynamic annotations on tensors in ListVariables/TupleVariables (#152119)

  Show description for 7aebb12
  

  xmfan

  authored and

  pytorchmergebot

  committed
  7aebb12

  Copy full SHA for 7aebb12

• [ca] hide unused scalar int sizes from dynamo (#151962)

  Show description for 4555ed8
  

  xmfan

  authored and

  pytorchmergebot

  committed
  4555ed8

  Copy full SHA for 4555ed8

• [ca] mark scalar int sizes as dynamic via tensor wrapping (#151731)

  Show description for 18229a5
  

  xmfan

  authored and

  pytorchmergebot

  committed
  18229a5

  Copy full SHA for 18229a5

• [aot][ca] save bw_module in AOTAutogradCache (#151860)

  Show description for 613bd46
  

  xmfan

  authored and

  pytorchmergebot

  committed
  613bd46

  Copy full SHA for 613bd46

• [aot] mark dynamic activations as maybe dynamic (#149707)

  Show description for c461ba6
  

  xmfan

  authored and

  pytorchmergebot

  committed
  c461ba6

  Copy full SHA for c461ba6

• BE: Swap functorch --> torch._higher_order_ops (#152620)

  Show description for b6c5886
  

  seemethere

  authored and

  pytorchmergebot

  committed
  b6c5886

  Copy full SHA for b6c5886

• Revert "[torchgen] Refactor `torchgen.utils.FileManager` to accept `pathlib.Path` (#150726)"

  Show description for 1c04ea4
  pytorchmergebot

  committed
  1c04ea4

  Copy full SHA for 1c04ea4

Pagination

Previous

Next