strisunshinewentingwang/pytorch

Pull Requests Commits

Cherry pick sorting patch (#85620)

shmsong committed 1 year ago

101f10d7

[MPS] Fix placeholder behavior for transposed view (#85689)

malfet committed 1 year ago

1367f240

Disallow auto_element_wise for in-place and fix some in-place gradients (#85634)

soulitzer committed 1 year ago

15c52ffc

Expose cpp_backtrace to python binding (#84896)

SherlockNoMad committed 1 year ago

01dbbeee

Don't use a fixed name to avoid race conditions. (#84952)

Rodrigo Kumpera committed 1 year ago

54e03cdd

Add __all__ to torch.utils submodules (#85331)

anjali411 committed 1 year ago

0183c1e3

resize_as_sparse support all compressed layouts (#85378)

amjames committed 1 year ago

f6485718

Optimize to if the datatyep of the source tensor is as same as the dest datatype (#85140)

EikanWang committed 1 year ago

45be74cc

Use high precision accmulate buffer for bf16 accmulation (#84402)

EikanWang committed 1 year ago

83261ff9

[vulkan] Rewrite prepacking functions using aten functions + some code cleanup (#84973)

SS-JIA committed 1 year ago

cf5699f2

Revert "Add environment parse function that supports default value (#85563)"

pytorchmergebot committed 1 year ago

b360d663

[torchdynamo hash update] update the pinned torchdynamo hash (#85683)

pytorchmergebot committed 1 year ago

e1e056ac

[vision hash update] update the pinned vision hash (#85684)

pytorchmergebot committed 1 year ago

8125d2e1

[MPS] Clamp op - fix shape issues (#114) (#85673)

abhudev committed 1 year ago

7a5449f1

Fix warning in kineto_shim.h (#85653)

r-barnes committed 1 year ago

cce6d8d6

[Modes] remove enable and rewrite mode stack (squashed) (#84774)

samdow committed 1 year ago

18d8c548

[MPS] Fix test consistency error 'mlir module expected element type ui8 but received si8' (#85666)

DenisVieriu97 committed 1 year ago

a0be0ca1

[MPS] Fix memory leaks that cause the buffers not to be released and cause OOM (#85661)

razarmehr committed 1 year ago

b8d2ab3d

[LRD] Allowing using dedicated iteration counter for learning rate (#85195)

Wenguang Mao committed 1 year ago

755b39ba

Add environment parse function that supports default value (#85563)

kwen2501 committed 1 year ago

784f4ba1

[maskedtensor] port torch/_masked into torch/masked (#85515)

george-qi committed 1 year ago

686555b6

Copy over non parameter grad (#85658)

eellison committed 1 year ago

90261945

Change API type `Tensor[]` for structured kernels. (#73350)

bdhirsh committed 1 year ago

4a2d2e5e

Fix broken periodic workflow after removing ios secret (#85664)

huydhn committed 1 year ago

1a2734e0

Enforce non-virtual-dtor everywhere (#85586)

huydhn committed 1 year ago

e4471032

[fx] Make NormalizeArgs preserve node type (#85637)

Mike Iovine committed 1 year ago

f325c29b

Match kwargs in SubgrpahMatcher (#85617)

SherlockNoMad committed 1 year ago

5547c6aa

Clean up the functorch test skip mechanism; add a new decorator (#85564)

zou3519 committed 1 year ago

e38b3424

Get flash_attn to compile for CUDA 11.6 linux nightly build (#84941)

cpuhrsch committed 1 year ago

6a04df3a

Configure PyTorch Testing ArgumentParser Instance To Avoid Unnecessary Conflicts with System Args (#85616)

speediedan committed 1 year ago

15435325

Newer Older