strisunshinewentingwang/pytorch

Pull Requests Commits

[inductor] fix bitwise codegen

Valentine233 committed 2 years ago

9b162780

Inductor cpp wrapper: support LinearUnary (#97655)

chunyuan-w committed 2 years ago

d03003ab

Inductor cpp wrapper: support MKLPackedLinear (#90755)

chunyuan-w committed 2 years ago

0c1f524b

[Inductor] support transpose vertical reduction in cpp (#97781)

Jiong Gong committed 2 years ago

5d62d125

Improve support for dict subclasses (#98154)

jansel committed 2 years ago

76074dc0

[Inductor] support vertical reduction in cpp (#97644)

Jiong Gong committed 2 years ago

bf22ecba

Fix graph break from inline patched init (#98150)

jansel committed 2 years ago

35b33095

[Inductor] simplify CPP backend Tile2D code and support non-contiguous load/store (#97626)

Jiong Gong committed 2 years ago

8e5f4916

[inductor] Fallback on complex64 kernels (#98155)

jansel committed 2 years ago

71d850a1

Support inlining no_grad() decorator (#98121)

jansel committed 2 years ago

bc9dd969

[Easy] Fix lint error on DTensor math_ops.py (#98170)

mrshenli committed 2 years ago

96403cfc

[Easy] Include SPMD and DTensor files in UFMT checks (#98148)

mrshenli committed 2 years ago

02179827

TensorExpr eval: fix copying variables from pointers on big endian systems (#96951)

AlekseiNikiforovIBM committed 2 years ago

38609cc4

Update ideep submodule (#97430)

yanbing-j committed 2 years ago

2ab18a23

[Easy] Consolidate string startswith checks (#98147)

mrshenli committed 2 years ago

347c67d4

[reland] switch mean to use reduction linear (#97996)

wanchaol committed 2 years ago

7fcff01b

Fix graph break from 'hasattr: HFPretrainedConfigVariable()' (#98119)

jansel committed 2 years ago

d9e5ab46

Improve support for contextlib.nullcontext (#98111)

jansel committed 2 years ago

b9d3b3f5

Add --stats option to benchmark scripts (#98109)

jansel committed 2 years ago

92b46202

avoid warning in irange for unsigned types (#97973)

mikey dagitses committed 2 years ago

e402259b

`masked_scatter` should accept only bool masks (#97999)

malfet committed 2 years ago

2af09393

Move CPUReproTests to its own file (#97943)

jansel committed 2 years ago

bbc4e911

[MPS] Enable conditional indexing tests (#97871)

qqaatw committed 2 years ago

db8abde9

[SPMD] Enable fused Adam in full train step tracing (#98113)

mrshenli committed 2 years ago

e8d39606

Format DTensor dispatch.py and _meta_registrations.py (#98114)

mrshenli committed 2 years ago

bccf2ef0

remove redundant typed StorageImpl::data() member (#97650)

mikey dagitses committed 2 years ago

64077ce5

[inductor] more cuda metrics in wrapper (#97723)

shunting314 committed 2 years ago

13461e97

[quant][pt2e][refactor] Remove extra arguments of _maybe_insert_observers_before_graph_output (#98029)

jerryzh168 committed 2 years ago

553bb01d

Call to mkldnn_matmul from aten::addmm on AArch64 (#91763)

milpuz01 committed 2 years ago

26301447

[vision hash update] update the pinned vision hash (#98108)

pytorchmergebot committed 2 years ago

57c6f3fe

Older