strisunshinewentingwang/pytorch

Pull Requests Commits

[ci] Move sm86 from trunk to pull (#93085)

clee2000 committed 1 year ago

ae171cf6

Fix topk IMA (#93095)

zasdfgbnm committed 1 year ago

d1807dc1

Make __deepcopy__ of GraphModule able to handle circular reference. (#93038)

qihqi committed 1 year ago

8d7f9e2f

[CI] Move parallel native builds to 3.8 (#93103)

malfet committed 1 year ago

ceb44350

[Reland] aot autograd explicitly errors on double backward (#92893)

soulitzer committed 1 year ago

f6f46ba3

Revert "Disable torch_jit_fuser_te for dynamo CI (#92945)"

pytorchmergebot committed 1 year ago

913cf290

Torchinductor randn_like lowering (#93005)

eellison committed 1 year ago

340811bf

Properly compute device for elementwise operations with CPU scalar tensor (#93073)

ezyang committed 1 year ago

1b5bfe9d

Update flatbuffer test models to match pkl models (#93022)

qihqi committed 1 year ago

1f352f7c

[MacOS] Explicitly use cmake from cloned conda environment (#92737)

huydhn committed 1 year ago

68a49322

Remove try catch in test_torchinductor (#93004)

eellison committed 1 year ago

15c46eb8

Make meshgrid support symbolic shapes (#93075)

ezyang committed 1 year ago

17803fb3

Don't copy name_to_input in OutputGraph (#93034)

mlazos committed 1 year ago

5de19dd3

Update XLA docker image to v0.8 (#93041)

huydhn committed 1 year ago

f30787e5

Update RELEASE.md with pinning xla and builder PRs (#93079)

atalman committed 1 year ago

d9f0d148

Add sparse COO tensor support to torch.sum(dim=..., keepdim=...) (#92979)

pearu committed 1 year ago

0e92bbe5

[BE][CI] Move more builds from 3.7 to 3.8 (#92928)

malfet committed 1 year ago

ca2a23c2

Setup shebang and set -x on generated runner script (#93007)

ezyang committed 1 year ago

729f1a8e

Revert "Improve `bsr @ strided` performance in `baddmm` for `bfloat16/half` with Triton kernels. (#88078)"

pytorchmergebot committed 1 year ago

7012d985

Apply some more missing moves in aten native (#92983)

Skylion007 committed 1 year ago

3888555f

Fix some silly Inductor bugs (#92997)

ezyang committed 1 year ago

7e449e8b

Revert spurious submodule change from #92107 (#93067)

ezyang committed 1 year ago

abcaa05f

Mark crossvit_9_240 as passing dynamic=True (#92981)

ezyang committed 1 year ago

5e9fa0a8

[Quant][Fx] Fix issue: qconfig_mappings of onednn backend are not correctly set for fused modules (#91297)

Xia-Weiwen committed 1 year ago

1d03a6a9

[PT-D][TP] Fix TP API for FQN path based parallelization (#93029)

fduwjj committed 1 year ago

913866ef

Improve `bsr @ strided` performance in `baddmm` for `bfloat16/half` with Triton kernels. (#88078)

nikitaved committed 1 year ago

46f16b93

[functorch][reland] vmap: bitwise operators (#92836)

khushi-411 committed 1 year ago

4c074ddf

Include cublasLt as an option in max_autotune mode (#92915)

jansel committed 1 year ago

ccad2e50

[functorch] fix batching rule for dropout (#92975)

kshitij12345 committed 1 year ago

d88bc38b

[PT-D] Enable Meta Tensor Support for DTensor (#92652)

fduwjj committed 1 year ago

77f33660

Newer Older