strisunshinewentingwang/pytorch

Pull Requests Commits

Remove unnecessary items() call in zero_grad (#97040)

Skylion007 committed 1 year ago

dd9ade63

[SDPA] Remove the chunk_grad from mem-eff attention (#96880)

drisspg committed 1 year ago

98a5cf09

Fail fast when dynamo attempts to add unspecialized int/float as additional graph inputs (#96786)

guangy10 committed 1 year ago

d4b8ed2b

Improve size mismatch error messaging referencing mat/vet sizes (#96863)

kiersten-stokes committed 1 year ago

cea13ad9

Bind increment_version to python (#96852)

albanD committed 1 year ago

985fc66b

Fixed print tensor.type() issue. (#96381)

NmomoN committed 1 year ago

1983b317

[static-runtime] one more attempt to improve crash log readability (#96903)

tenpercent committed 1 year ago

57bb5b15

[cpp extension] Allow setting PYTORCH_NVCC to a customized nvcc in torch cpp extension build (#96987)

xwang233 committed 1 year ago

44d7bbfe

[ez][inductor] show kernel category in kernel benchmark result (#96991)

shunting314 committed 1 year ago

8ce296ae

Fix Typo in pytorch/torch/autograd/__init__.py (#97024)

LukeConfait committed 1 year ago

46eaf4be

[DTensor] Fix _get_or_create_default_group() (#96961)

shintaro-iwasaki committed 1 year ago

95575f0a

Change `THPStorage::cdata` to be a `MaybeOwned<Storage>`, add unpack func (#96801)

kurtamohler committed 1 year ago

ffddb221

test/test_torch.py: fix TestTorch::test_from_buffer test (#96952)

AlekseiNikiforovIBM committed 1 year ago

7f94ea84

[Inductor] preserve AliasedLayout on View (#96948)

jgong5 committed 1 year ago

18cf30fb

Decoder native functions join the dead code society (#96025)

mikekgfb committed 1 year ago

92eb9d36

Revert "[aot autograd] refactor to make functionalization self-contained (#96341)"

pytorchmergebot committed 1 year ago

b5ecf727

inductor: fix cpp wrapper ExternKernel check (#96799)

chunyuan-w committed 1 year ago

238b0608

[1/n] Consolidate `replicate` and `DDP`: setup ufmt for `distributed.py` (#96597)

yhcharles committed 1 year ago

13538c88

Move `hasPrimaryContext` to `c10::cuda` (#96800)

malfet committed 1 year ago

24ce3a7c

[vision hash update] update the pinned vision hash (#96990)

pytorchmergebot committed 1 year ago

cbd3df93

[PyTorch][XNNPACK] Update wrappers for internal only x86 SSE2 kernels (#96896)

SS-JIA committed 1 year ago

4de1bc16

[MPS] Introduce MPSUnaryGradCachedGraph & MPSBinaryGradCachedGraph (#95289)

qqaatw committed 1 year ago

f865e23a

cudagraph trees (#89146)

eellison committed 1 year ago

571f96bf

nn.EmbeddingBag bound check (#96022)

xw285cornell committed 1 year ago

cf732053

[MPS] Fix the failure with ReplicatePad3D (#96988)

razarmehr committed 1 year ago

50beab29

Revert "[PTD][Checkpoint] Add checkpointing support for DTensor submesh (#96802)"

pytorchmergebot committed 1 year ago

417e7bc0

Stop using my channel for 3.11 builds (#96973)

malfet committed 1 year ago

c9135e44

record caller frame instead of function frame (#96882)

avikchaudhuri committed 1 year ago

e4e761b2

Expose Stream Recording Apis in python (#96384)

eellison committed 1 year ago

ea741508

addmv decomp #2 (#96264)

cybershiptrooper committed 1 year ago

b01d6f2c

Newer Older