strisunshinewentingwang/pytorch

Pull Requests Commits

Skip TestNNAPI tests if QNNPACK is not supported (#82882)

Flamefire committed 2 years ago

cafdc52c

Make M1 tests green (#82213)

albanD committed 2 years ago

2255911f

Fix leak when create_graph and full backward hook registered (#82788)

soulitzer committed 2 years ago

1cafb102

Move OpInfo subclasses into opinfo.core (#82830)

peterbell10 committed 2 years ago

d8ae83ba

Move OpInfo class into new opinfo folder (#82540)

peterbell10 committed 2 years ago

4d405517

[Quant][fx] Remove dequant-quant around getitem (#82675)

andrewor14 committed 2 years ago

8f38f677

[vulkan] Pass correct output_padding_arg when creating conv2d contexts (#82861)

SS-JIA committed 2 years ago

18858765

Handle complex optimization in AdamW by treating complex numbers as 2D real numbers (#80280)

zaxtax committed 2 years ago

dcbe9ce2

[xla hash update] update the pinned xla hash (#82877)

pytorchmergebot committed 2 years ago

0f32d71b

[torchdynamo hash update] update the pinned torchdynamo hash (#82876)

pytorchmergebot committed 2 years ago

7b486667

Add DLPack support for XPU backend by mapping to kDLOneAPI in DLPack (#82867)

chengjunlu committed 2 years ago

de0e0300

Re-land sym_numel (#82374) (#82726) (#82731) (#82855)

Krovatkin committed 2 years ago

bfebf254

Fix profiling with record_shapes=True and nested tensor (#82854)

ngimel committed 2 years ago

bdb0abb2

fix matching against MemcpyAsync (#82782)

davidchencsl committed 2 years ago

e47637aa

Migrate remaining pytorch code to use new flatbuffer_loader.h APIs (#82620)

dbort committed 2 years ago

6e712823

Remove flatbuffer types/headers from flatbuffer_serializer[_jit].h (#82619)

dbort committed 2 years ago

0810961d

New flatbuffer_loader functions that do not depend on flatbuffers.h (#82618)

dbort committed 2 years ago

802a4fd2

Back out "Add an op_lowering_disallow_list in fx splitter base class. (#82288)" (#82750)

Shirong Wu committed 2 years ago

4ae40d74

Revert "Add DLPack support for XPU backend by mapping to kDLOneAPI in DLPack. (#81021)"

pytorchmergebot committed 2 years ago

0e16340f

Revert "Fix possible out-of-range error of 'as_nd' util function (#82393)"

pytorchmergebot committed 2 years ago

dc788f6d

[ONNX] Add RReLU eval mode behavior (#82678)

titaiwangms committed 2 years ago

b18498a6

Add a utility function for quantizing bias (#82848)

taivu1998 committed 2 years ago

c434bc2f

Fix possible out-of-range error of 'as_nd' util function (#82393)

Xia-Weiwen committed 2 years ago

1ac0285c

[NNC] Skip buildShapeExpressions if ConstantChunk input shapes are unknown (#82698)

davidberard98 committed 2 years ago

b721ea9b

[NestedTensor]Remove tensor buffer replace with Storage type (#82757)

drisspg committed 2 years ago

2f9d046d

Test that multi_tensor optimizer state buffers match with single_tensor state buffers (#81894)

mikaylagawarecki committed 2 years ago

b1922e03

Revert "Re-land sym_numel (#82374) (#82726) (#82731)"

pytorchmergebot committed 2 years ago

78bd95b1

[functorch] Add a bunch of low hanging fruit linalg batch rules (#82177)

samdow committed 2 years ago

ae399d00

Refactor quant levels visualization (#82790)

asl3 committed 2 years ago

34103a30

Fix a segfault in `new_empty_strided` (#82422)

Nikita Karetnikov committed 2 years ago

f6749e76

Newer Older