strisunshinewentingwang/pytorch

Pull Requests Commits

[ONNX] Support type promoting sym number representing scalar output (#106178)

BowenBao committed 1 year ago

5c3aae83

[memory snaphots] record for all devices (#106346)

zdevito committed 1 year ago

449f481d

[DCP][Test]Remove broken 2d checkpoint test (#106367)

wz337 committed 1 year ago

d2a9b256

Modify signature for tensor.tile in doc (#106295)

jennychristopher committed 1 year ago

4b2c6282

[ONNX] Do not run 'deduplicate_initializers' when 'keep_initializers_as_inputs' is True (#96320)

BowenBao committed 1 year ago

05b2a6c8

[SDPA] Update dispatch checks to catch last_dim_stride != 1. Also update mask padding logic (#106102)

drisspg committed 1 year ago

cfa4edcd

[Doc] Add proj_size < hidden_size in LSTM (#106364)

isdanni committed 1 year ago

1a6f1d81

Remove fake_mode arg from torch._dynamo.export API (#106345)

Thiago Crepaldi committed 1 year ago

6d2162e6

Propagate dynamic int on `__setitem__`. (#105923)

ysiraichi committed 1 year ago

596491f1

Do not call decref if python runtime is already dead (#106334)

albanD committed 1 year ago

cf012c43

Revert "Add scalar conversion using avx instructions for half (#102140)"

pytorchmergebot committed 1 year ago

6d86a255

[Export] remove unused flags in export (#106336)

ydwu4 committed 1 year ago

aaaafa1b

[caffe2] Move CAFFE SDT macros' definitions to `c10/util/` (#105856)

vlad-scherbich committed 1 year ago

e35950cd

[inductor] Fix lowerings that create unexpected aliases (#105173)

peterbell10 committed 1 year ago

4dc06382

Normalize builtin types to dtypes. (#106074)

ysiraichi committed 1 year ago

e5143863

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

mikaylagawarecki committed 1 year ago

87d25369

Add error checking for padding modules (#106147)

mikaylagawarecki committed 1 year ago

0547b627

Add error inputs to ModuleInfo (mirroring OpInfo) (#106325)

mikaylagawarecki committed 1 year ago

c9be60cd

remove tensorpipe code which forgot to delete (#106301)

fakeYan committed 1 year ago

16df5423

[pt2] add meta for `ormqr` (#106278)

nkaretnikov committed 1 year ago

f23d755e

[opinfo] Fix logic in sample_inputs_linspace (#106353)

justinchuby committed 1 year ago

780b90ba

Only make a shallow copy when loading optimizer state_dict (#106082)

janeyx99 committed 1 year ago

59d0dea9

[vision hash update] update the pinned vision hash (#106350)

pytorchupdatebot committed 1 year ago

ceea08a9

[Pytorch][Vulkan] Reuse broadcast checks (#105960)

lucylq committed 1 year ago

aa2cee44

Update kineto submodule to 465ff4cd7 (#106293)

davidberard98 committed 1 year ago

f456c504

create benchmark example tensors with correct sizes (#106238)

chenyang78 committed 1 year ago

f1109028

[Inductor][FX passes] New group/batch fusion pattern searching algorithm + group mm fusion + preserve memory layout (#106279)

yanboliang committed 1 year ago

03e85be9

Update Anaconda download link (#106335)

dzienisz committed 1 year ago

a2e8ac1d

Extend workflow sync to more workflow (#106331)

malfet committed 1 year ago

e075f91d

fix sdpa math accuracy issue when scale is negative (#105202)

XiaobingSuper committed 1 year ago

55f9359d

Newer Older