strisunshinewentingwang/pytorch

Pull Requests Commits

alanwaketan committed 1 year ago

36043be0

update xla hash

alanwaketan committed 1 year ago

b2f8b80f

[LTC] Make LazyGraphExecutor::RunPostOrder() virtual

alanwaketan committed 1 year ago

7fe9693d

Add mkl implementation for exponential on CPU (#69967)

CaoE committed 1 year ago

eae0f3f5

[LTC] Make even more LazyGraphExecutor interfaces virtual (#90650)

alanwaketan committed 1 year ago

a50fe978

[FSDP] Clean up `FlatParamHandle` dtypes, post-backward hook (#90660)

awgu committed 1 year ago

fc429512

TorchDynamo: always convert tensor to fake tensor at fake_mode path for ShapeProp (#90685)

XiaobingSuper committed 1 year ago

ffa89033

Remove hard numpy dep introduced by _inductor/utils.py (#90716)

amjames committed 1 year ago

7a7f2970

Upgrade CI to ROCm5.3 (#88297)

jithunnair-amd committed 1 year ago

181a82ff

Re-enabled 2 Metaprogramming tests on Windows (#87284)

mantaionut committed 1 year ago

7498e23b

Remove hack to hard code test times (#90720)

soulitzer committed 1 year ago

dc4d18d4

[c10d] remove some outdated bc checks for c10d op (#90681)

wanchaol committed 1 year ago

1f86a144

[c10d] update alltoall signature to be more consistent (#90569)

wanchaol committed 1 year ago

7da50450

Add allgather_into_tensor to CommTensor (#90565)

wanchaol committed 1 year ago

f30694c7

Add reduce_scatter_tensor to CommTensor (#90564)

wanchaol committed 1 year ago

b782927e

Add alltoall_ to CommTensor (#90512)

wanchaol committed 1 year ago

3ba9e4cd

[LTC] Make DeviceContextArena protected (#90531)

alanwaketan committed 1 year ago

6165a180

Guard Symbol and ShapeGuardPrinter behind HAS_SYMPY (#90704)

seemethere committed 1 year ago

b8f35ec6

Revert "[torchgen] Let native function declaration generation logic take a callable (#90590)"

pytorchmergebot committed 1 year ago

ea64c8c6

Revert "[torchgen] Introduce Executorch types and signatures (#90591)"

pytorchmergebot committed 1 year ago

b3e6a6dc

Create stub function for doing SDPA cpp and cuda dispatch (#90576)

drisspg committed 1 year ago

42a5f6ee

Fix non-existing parameters in docstrings in torch/fx (#90594)

kit1980 committed 1 year ago

df569367

[quant] Add example for lowering quantized dynamic linear pattern through delegation (#90640)

jerryzh168 committed 1 year ago

94b9bb32

Fix a minor typo in documentation (#90667)

kevmo314 committed 1 year ago

b6f114c2

Fix prelu ref when a.ndim < 2 (#89809)

soulitzer committed 1 year ago

98a9235d

Add comment to output_code in dynamo config (#90333)

williamwen42 committed 1 year ago

34dc34e8

move TypedStorage handling to assertEqual (#89557)

pmeier committed 1 year ago

7bb97c4c

Fix a typo in some torch.load error message. (#90662)

LaurentMazare committed 1 year ago

17941b12

[Dynamo] Supports calling parent class‘s non classmethod from child class (#90682)

yanboliang committed 1 year ago

e2674aaf

[ao] fix incorrect integer cast on histogram observer bounds (#90355)

HDCharles committed 1 year ago

e1165088

Older