strisunshinewentingwang/pytorch

Pull Requests Commits

create API jit::Module::deepcopy(device) (#106521)

jiayisuse committed 1 year ago

8ef7512d

_force_original_view_tracking to work as both context manager and function (#106706)

poseljacob committed 1 year ago

a25eee1d

[inductor] Fix xpasses being impossible (#106631)

peterbell10 committed 1 year ago

3cda19c1

[pt2] Add reference implementations of torch.{stft,istft} (#106400)

peterbell10 committed 1 year ago

ab6efb16

[FakeTensor] Workaround FFT ops with incorrect meta strides (#106319)

peterbell10 committed 1 year ago

d4d090e2

[inductor][fx passes] add tensor size limit for group fusion and enable batch fusion (#106627)

jackiexu1992 committed 1 year ago

66d90e80

Better dynamo dict support via SetVariable keys (#106559)

voznesenskym committed 1 year ago

45c03b1a

[BE] Use convenience function (#106709)

malfet committed 1 year ago

4639ceb3

[BE] Use nested namespace in `ATen/cuda` (#106708)

malfet committed 1 year ago

e02a3d4a

Reland "Add nn.CircularPad{*}d for consistency + fix no_batch_dim support (#106148)" (#106632)

mikaylagawarecki committed 1 year ago

1317dbf1

[NAdam] Add capturable API and tests + fix differentiable (#106615)

janeyx99 committed 1 year ago

0208574d

[ROCM] enable test_aten cpp tests (#106476)

hongxiayang committed 1 year ago

3dd8cb12

Fix lifetime of JITException binding (#106401)

albanD committed 1 year ago

3a07dfde

[functorch] fix dynamo support for functorch.grad (#106610)

kshitij12345 committed 1 year ago

af78e139

[Fix Bug] Cannot assign index like x[[1,2], :] = 2 when torch.use_deterministic_algorithms(True) to main (#105833)

wey-code committed 1 year ago

2d4b1ae4

Handle `Rational` divisors in `FloorDiv`. (#106644)

ysiraichi committed 1 year ago

070eb88a

More readable Z3 expressions printer. (#106643)

ysiraichi committed 1 year ago

33e70e34

export tools/autograd to torchgen package (#106663)

heidongxianhua committed 1 year ago

26846546

Fix typo in jit_opt_limit.h (#106684)

huihoaan committed 1 year ago

273ad1dd

[xla hash update] update the pinned xla hash (#106695)

pytorchupdatebot committed 1 year ago

1cc00262

[Inductor][FX passes] Pre grad batch linear LHS fusion (#106497)

yanboliang committed 1 year ago

416bf4e3

[DCP][Test]Remove broken 2d checkpoint test (#106640)

wz337 committed 1 year ago

e35cb480

fix `upsample_nearest` decompositions for `uint8` tensors (#106675)

ar0ck committed 1 year ago

aa1b2f16

conv-bn folding in low precision (#106576)

eellison committed 1 year ago

c21df02e

[pt2] add Python meta for `polygamma` (#106681)

nkaretnikov committed 1 year ago

7215007f

Use default dispatch table for `tensordot.out` (#106669)

nkaretnikov committed 1 year ago

12041d8e

[pt2] add meta for `_cdist_backward` (#106680)

nkaretnikov committed 1 year ago

f694bcc9

[pt2] remove meta skips for `aminmax`, decomp exists (#106670)

nkaretnikov committed 1 year ago

05e1a507

Improve AOTAutograd tests to do something when inputs don't require grad (#106558)

zou3519 committed 1 year ago

26e98040

Minor update to ROCm triton commit pin (#106616)

jataylo committed 1 year ago

8b8f576f

Newer Older