strisunshinewentingwang/pytorch

Pull Requests Commits

[Reland] Refactor stack_trace preservation for node meta preservation (#90803) (#90803)

SherlockNoMad committed 2 years ago

37c03e37

refactor: move dynamo/TorchXLA bridge to pytorch/xla repo (#92601)

shunting314 committed 2 years ago

67689c82

[Py3.11] Remove skip logic from vmap and forward_ad (#91825)

malfet committed 2 years ago

b2f3ff61

Apply some std::move and param value fixups to aten (#92901)

Skylion007 committed 2 years ago

f2f42e54

Added keep_key option to Grouper (#92532)

SvenDS9 committed 2 years ago

b073c09f

Add --timing and --explain to CI runs (#92980)

ezyang committed 2 years ago

63331a5f

[cpp] remove checks from embedding bag impl (#92982)

suo committed 2 years ago

63e47c68

Disable vml's abs and log1p (#92113)

peterbell10 committed 2 years ago

99ced648

Improve performance for unary kernels using vml (#91963)

peterbell10 committed 2 years ago

d4c8e37b

Add get-job-id in get-workflow-job-id action (#93001)

clee2000 committed 2 years ago

0de81906

adding some more missing ops to vmap (#92110)

srossross committed 2 years ago

d354499f

Upload failures shouldn't fail a CI that passed tests (#92996)

ZainRizvi committed 2 years ago

92fbb35b

More clang-tidy fixes (#92944)

cyyever committed 2 years ago

e292ddff

Add few more tests to 3.11 smokechecks (#92946)

malfet committed 2 years ago

4e673326

Make TensorIterator give better error message for symbolic tensors (#92914)

ezyang committed 2 years ago

b399007a

[FSDP] Fix `no_sync()`, `use_orig_params=True`, mixed precision, sharded (#92874)

Andrew Gu committed 2 years ago

c0ed0f22

add number of cuda retries into tracker (#92557)

zhaojuanmao committed 2 years ago

077e135e

Rename Canonical Aten IR to Core Aten IR (#92904)

SherlockNoMad committed 2 years ago

a6ac922e

Fix to use upsample_bicubic2d.vec decomp for dynamic shape support (#92854)

jbschlosser committed 2 years ago

e5fd7e6d

Disable torch_jit_fuser_te for dynamo CI (#92945)

malfet committed 2 years ago

0fc2f9fe

Change ciflow/inductor to test inductor inference with dynamic shapes (#92771)

ezyang committed 2 years ago

2ee94633

Add dynamic shapes aot_eager to periodic (#92770)

ezyang committed 2 years ago

f724ecbd

Fix #92814: assertion error when explicitly provide out=None (#92873)

ezyang committed 2 years ago

9c487a4b

Revert "[CI] Disable regularly failing CUDA 11.8 windows periodic tests (#92902)"

pytorchmergebot committed 2 years ago

f180873f

[inductor] skip CUDA tests under ASAN (#92883)

nkaretnikov committed 2 years ago

e45b5660

Remove windows check for cmake to build Fused kernels (#91909)

drisspg committed 2 years ago

a3715efd

[CI] Rename TSAN job (#92929)

malfet committed 2 years ago

f0d09572

Revert "Fix to use upsample_bicubic2d.vec decomp for dynamic shape support (#92854)"

pytorchmergebot committed 2 years ago

01f10977

lru_cache shape expansion (20-25% speedup on local bench) (#92860)

voznesenskym committed 2 years ago

54bbb446

[dynamo] Allow DynamicShapeVariable as predicate to cond() op. (#92864)

zhxchen17 committed 2 years ago

78caa792

Older