strisunshinewentingwang/pytorch

Pull Requests Commits

`_foreach_copy` with different src/dst dtypes (#121717)

crcrpar committed 1 year ago

da2a9a05

[dynamo] Minor compile time optimizations in torch.py (#121615)

jansel committed 1 year ago

a13dd92d

[executorch hash update] update the pinned executorch hash (#121056)

pytorchupdatebot committed 1 year ago

d619be57

CI: Fix flaky artifact upload step (#121733)

peterbell10 committed 1 year ago

0c1d59b7

[inductor] Update triton pin (#121268)

peterbell10 committed 1 year ago

52ed35bb

[MPS][BE] Define `_compute_tolerances` (#121754)

malfet committed 1 year ago

07330ff7

cublasLt workspace warning info is misleading, the unit of measuremen… (#121073)

whitelok committed 1 year ago

f83392b6

[ROCm] Enable several test_unary_ufuncs UTs on ROCm (#121104)

BLOrange-AMD committed 1 year ago

e755dab0

[AOTInductor] Skip tests on RoCM for duplicate_constant_folding (#121750)

muchulee8 committed 1 year ago

f24ae66a

Gate tt.reduce Triton mutation tests on Triton version (#121753)

aakhundov committed 1 year ago

9f235971

Remove error anti-pattern when dealing with dynamic shape output (#121681)

gmagogsfm committed 1 year ago

7d05c4c0

Revert "[ Inductor ] Shape padding honors output stride preservation (#120797)"

pytorchmergebot committed 1 year ago

9df0dca7

[torch export] replace traceback.extract_stack with CapturedTraceback.extract (#121449)

strisunshinewentingwang committed 1 year ago

02bb2180

CI: Specify libc and libstdcxx versions in conda environments (#121556)

peterbell10 committed 1 year ago

7a53dedb

Cleanup some exception handling in triton mutation tracking (#121739)

oulgen committed 1 year ago

68be750e

Fix aoti doc to avoid cannot bind non-const lvalue reference error (#121672)

mreso committed 1 year ago

a9274c9a

Support `triton.language.dtype` with `torch.compile` (#121690)

oulgen committed 1 year ago

79ee6bbd

[dynamo][guards] Use lazy variable tracker for func defaults (#121388)

anijain2305 committed 1 year ago

22bb2498

[fx] Preserve Fx graph node order in partitioner across runs (#115621)

kareemshaik80 committed 1 year ago

519151a0

Release fix pinning slow-tests.json (#121746)

atalman committed 1 year ago

a95ceb51

[Inductor Cutlass backend] Move tests to separate file (#121489)

kadeng committed 1 year ago

a5ec45f2

feat: Update Dockerfile default versions for Python, OS, and CUDA arch list (#121560)

bryantbiggs committed 1 year ago

844bfbbd

[Profiler] add missing field device_resource_id (#121480)

lucifer1004 committed 1 year ago

d62bdb08

Don't run non-strict for test case that doesn't need non-strict (#121710)

tugsbayasgalan committed 1 year ago

5478a4e3

Revert "[dynamo][guards] Use lazy variable tracker for func defaults (#121388)"

pytorchmergebot committed 1 year ago

5b506c8b

[eazy] add more log when accuracy check fail (#121656)

shunting314 committed 1 year ago

522d9729

avoid aten dispatch shadowing type with variable (#121659)

EscapeZero committed 1 year ago

f50c6524

[pytorch] optional zero points on dequantize per channel (#121724)

manuelcandales committed 1 year ago

6d8a7d6e

[easy] Refactor MultiOutput. codegen_list_tuple_access to use subclass type checks (#121662)

ColinPeppler committed 1 year ago

a6149eba

Sanity check for non-strict (#121687)

tugsbayasgalan committed 1 year ago

90e886aa

Newer Older