Commits

Breadcrumbs

History for

onmain

User selector

All users

Datepicker

All time

Commit History

Commits on Oct 3, 2024

• Merge branch 'pytorch:main' into main

  jakeharmon8

  authored
  db63884

  Copy full SHA for db63884

• Revert "Fix unlift to unblock training IR + run_decomp on aliasing constants (pytorch#137162)"

  Show description for 525f671
  pytorchmergebot

  committed
  525f671

  Copy full SHA for 525f671

• Fix unlift to unblock training IR + run_decomp on aliasing constants (pytorch#137162)

  Show description for f96020c
  

  tugsbayasgalan

  authored and

  pytorchmergebot

  committed
  f96020c

  Copy full SHA for f96020c

• Fix serialization for torch.uint16, torch.uint32, torch.uint64 (pytorch#137184)

  Show description for b6cb174
  

  mikaylagawarecki

  authored and

  pytorchmergebot

  committed
  b6cb174

  Copy full SHA for b6cb174

• cond_batch_rule with boolean pred (pytorch#135009)

  Show description for bd9517c
  

  vmoens

  authored and

  pytorchmergebot

  committed
  bd9517c

  Copy full SHA for bd9517c

• Revert "raw_alloc ignores PYTORCH_NO_CUDA_MEMORY_CACHING (pytorch#131114)"

  Show description for 0d1701f
  pytorchmergebot

  committed
  0d1701f

  Copy full SHA for 0d1701f

• Unify cpp_extension build directory removal (pytorch#136059)

  Show description for b86269f
  

  xmfan

  authored and

  pytorchmergebot

  committed
  b86269f

  Copy full SHA for b86269f

• [DTensor] Register replication strategy for a few upsampling interpolate ops (pytorch#137201)

  Show description for 55c343f
  

  wz337

  authored and

  pytorchmergebot

  committed
  55c343f

  Copy full SHA for 55c343f

• Skip Flaky Test: for pytorch#134602 (pytorch#137226)

  Show description for 5c0ce8d
  

  drisspg

  authored and

  pytorchmergebot

  committed
  5c0ce8d

  Copy full SHA for 5c0ce8d

• [Inductor] Use parametrize to break down some unit tests (pytorch#137156)

  Show description for 4513fb5
  

  desertfire

  authored and

  pytorchmergebot

  committed
  4513fb5

  Copy full SHA for 4513fb5

• unflatten with specialized graphs per submodule call (pytorch#137013)

  Show description for cd5d1fe
  

  avikchaudhuri

  authored and

  pytorchmergebot

  committed
  cd5d1fe

  Copy full SHA for cd5d1fe

Commits on Oct 2, 2024

• Revert "[FSDP2] support torch._foreach_copy_(float8) for fully_shard(Float8Linear) (pytorch#135955)"

  Show description for 2c9e194
  pytorchmergebot

  committed
  2c9e194

  Copy full SHA for 2c9e194

• [FlexAttention] Fix max-autotune when captured buffers are View nodes (pytorch#137204)

  Show description for bb03ef7
  

  drisspg

  authored and

  pytorchmergebot

  committed
  bb03ef7

  Copy full SHA for bb03ef7

• Fix NJT serialization (pytorch#137031)

  Show description for e95b230
  

  jbschlosser

  authored and

  pytorchmergebot

  committed
  e95b230

  Copy full SHA for e95b230

• [SDPA] Bump grad_query fudge factor for Flash Attention (pytorch#135711)

  Show description for be423a8
  

  eqy

  authored and

  pytorchmergebot

  committed
  be423a8

  Copy full SHA for be423a8

• Check for fused kernel before inplace update (pytorch#137042)

  Show description for 36fb342
  

  exclamaforte

  authored and

  pytorchmergebot

  committed
  36fb342

  Copy full SHA for 36fb342

• Make PT2E work with both IR simultaneously (pytorch#135769)

  Show description for a3f3773
  

  yushangdi

  authored and

  pytorchmergebot

  committed
  a3f3773

  Copy full SHA for a3f3773

• improve get_schedule_class() (pytorch#137103)

  Show description for 4a9225f
  

  H-Huang

  authored and

  pytorchmergebot

  committed
  4a9225f

  Copy full SHA for 4a9225f

• [FSDP2] support torch._foreach_copy_(float8) for fully_shard(Float8Linear) (pytorch#135955)

  Show description for b50b3b3
  

  weifengpy

  authored and

  pytorchmergebot

  committed
  b50b3b3

  Copy full SHA for b50b3b3

• [inductor mkldnn test][BE] Use parametrize to shorten test run time (pytorch#137153)

  Show description for c318baf
  

  henrylhtsang

  authored and

  pytorchmergebot

  committed
  c318baf

  Copy full SHA for c318baf

• raw_alloc ignores PYTORCH_NO_CUDA_MEMORY_CACHING (pytorch#131114)

  Show description for 7001907
  

  2 peopleauthored and

  pytorchmergebot

  committed
  7001907

  Copy full SHA for 7001907

• [Inductor] External callable registration API for Matmul tuning candidates (pytorch#130774)

  Show description for a954a9e
  

  2 peopleauthored and

  pytorchmergebot

  committed
  a954a9e

  Copy full SHA for a954a9e

• [dynamo][user-defined-class] Fallback when object.__new__ fails (pytorch#137044)

  Show description for af86a6f
  

  anijain2305

  authored and

  pytorchmergebot

  committed
  af86a6f

  Copy full SHA for af86a6f

• Use torch.Stream&torch.Event for Dynamo capature (pytorch#134850)

  Show description for d290948
  

  guangyey

  authored and

  pytorchmergebot

  committed
  d290948

  Copy full SHA for d290948

• dont let partitioner think it can fuse pointwise ops into user triton kernels (pytorch#136878)

  Show description for bf73af4
  

  bdhirsh

  authored and

  pytorchmergebot

  committed
  bf73af4

  Copy full SHA for bf73af4

• [3/3][Inductor] Make CK work in FBCode (pytorch#136234)

  Show description for d117ec1
  

  ColinPeppler

  authored and

  pytorchmergebot

  committed
  d117ec1

  Copy full SHA for d117ec1

• [distributed][test] Remove unused variable and fix doc typo (pytorch#136943)

  Show description for e374d68
  

  zeshengzong

  authored and

  pytorchmergebot

  committed
  e374d68

  Copy full SHA for e374d68

• Add compiled_autograd_kwargs_override Dynamo config (pytorch#136967)

  Show description for a89e3c2
  

  yf225

  authored and

  pytorchmergebot

  committed
  a89e3c2

  Copy full SHA for a89e3c2

• Fix rendezvous error due to EtcdStore get method not waiting in some cases (pytorch#137056)

  Show description for 18525e1
  

  2 peopleauthored and

  pytorchmergebot

  committed
  18525e1

  Copy full SHA for 18525e1

• [logging/debugging] handle None (constant) args in debug log (pytorch#137032)

  Show description for f108f88
  

  coconutruben

  authored and

  pytorchmergebot

  committed
  f108f88

  Copy full SHA for f108f88

• Ensure noncontiguous tensor creation tests offsetting (pytorch#136396)

  Show description for f984b88
  

  benjaminglass1

  authored and

  pytorchmergebot

  committed
  f984b88

  Copy full SHA for f984b88

• fix build error with gcc 12+ (pytorch#137092)

  Show description for 63d6908
  

  abhishek-iitmadras

  authored and

  pytorchmergebot

  committed
  63d6908

  Copy full SHA for 63d6908

• [ts_converter] Fix prim::If buffer names (pytorch#136648)

  Show description for d725758
  

  angelayi

  authored and

  pytorchmergebot

  committed
  d725758

  Copy full SHA for d725758

Commits on Oct 1, 2024

• [numpy2.0 compat] Fix test_parse_numpy_int_overflow for NumPy 2.0

  Show description for bc4a8f9
  jakeharmon8

  committed
  bc4a8f9

  Copy full SHA for bc4a8f9

• Update incompatible cudagraph ops skip message (pytorch#137015)

  Show description for 0788d01
  

  eellison

  authored and

  pytorchmergebot

  committed
  0788d01

  Copy full SHA for 0788d01

Pagination

Previous

Next