strisunshinewentingwang/pytorch

Pull Requests Commits

Update cuda.rst

eqy committed 2 years ago

Verified 8b743c34

[BE] Update notes linkage in common_device_type, fix very minor grammar (#100898)

janeyx99 committed 2 years ago

9a811d1d

[3/n] loading meta to device (#100495)

qxy11 committed 2 years ago

812cadf9

[S337714] Back out "[PyTorch] Don't do extra numel() check in TensorImpl::data() (#98090)" (#100893)

xunnanxu committed 2 years ago

bde7b81f

[export] Fix cond for pass_base (#100836)

angelayi committed 2 years ago

2d2f716d

fix specify_constraints's signature when exporting model (#100739)

ydwu4 committed 2 years ago

b0a372e1

Forward fix lint failure from #100661 (#100907)

huydhn committed 2 years ago

fb69aa15

Always run prioritized tests first, even if they're expected to run serially (#100748)

ZainRizvi committed 2 years ago

95f191a2

[Dynamo] Remove cross import in dynamo unit tests (#100851)

yanboliang committed 2 years ago

c4bbeb5b

[inductor] Add variable names to MemoryDep (#100308)

jansel committed 2 years ago

5079bf3d

Fix nightly build of C++ docs (#100845)

huydhn committed 2 years ago

651b5b0f

[MPS] Add support for Custom Kernels (#100661)

razarmehr committed 2 years ago

f39cda83

Skip DNS host fix on ROCm runners (#100861)

huydhn committed 2 years ago

d9d98b4d

remove double doc upload after CloudFront fix (#99032)

pmeier committed 2 years ago

aaa1323c

Initialize view_replay_enabled_ in the AutogradState ctor (#100822)

vors committed 2 years ago

116e04be

handle new param from torch.compile (Inductor pattern matcher), enable_log (#100814)

lessw2020 committed 2 years ago

ec144b94

[stronghold][bc-linter] Switch to reusable action, enable for everyone (#100737)

izaitsevfb committed 2 years ago

ccd060ab

[inductor] Fix bug where a node gets erased twice (#100848)

jansel committed 2 years ago

176ef97f

[PyTorch/Distributed]Only sync buffers when broadcast_buffers is True (#100729)

xing-liu committed 2 years ago

07314206

[WIP] enable cuda graphs support for flash attention with dropout (#100196)

Natalia Gimelshein committed 2 years ago

bfe5f5bb

`USE_PRECOMPILED_HEADERS` is not supported on Apple M1 (#92868)

TachikakaMin committed 2 years ago

bb28f3f5

[inductor] Move cpp wrapper trigger logic to inner_compile (#100611)

desertfire committed 2 years ago

86ddfc7f

Revert "[xla hash update] update the pinned xla hash (#100369)"

pytorchmergebot committed 2 years ago

62c53aab

[pt2] add `SymInt` support for `linalg_qr_backward` (#100833)

nkaretnikov committed 2 years ago

6eb0d754

[pt2] add meta function for `solve_triangular` (#100829)

nkaretnikov committed 2 years ago

1e591a8b

bsr_dense_mm(): code refactoring (#100634)

nikitaved committed 2 years ago

cd8b82e5

[xla hash update] update the pinned xla hash (#100369)

pytorchmergebot committed 2 years ago

41bafb0b

Fix some NVCC warnings (#100823)

cyyever committed 2 years ago

333de1fd

inductor(CPU): skip weight packing when autocast is enabled (#100844)

XiaobingSuper committed 2 years ago

46affcb0

inductor(CPU): don't do binary fusion if binary's inputs are same tensor (#100843)

XiaobingSuper committed 2 years ago

92cecb8e

Older