strisunshinewentingwang/pytorch

Pull Requests Commits

extract out //c10/util:base library (#70854)

Michael Dagitses committed 2 years ago

40e88b75

[Array API] Add linalg.diagonal (#70599)

lezcano committed 2 years ago

108b37db

[jit][edge] Migrate to TypeFactory for jit types on mobile (#71516)

zhxchen17 committed 2 years ago

fe277b87

[acc_tracer] Do not rewrite the leaf modules (#71790)

Yinghai Lu committed 2 years ago

e5794974

[Easy] DDP typo fix (#71607)

rohan-varma committed 2 years ago

d3354602

[Opt Overlap] Implement register_fused_optim in DDP (#71606)

rohan-varma committed 2 years ago

10ca760c

[Opt Overlap] Implement _OverlappedOptimizer (#71605)

rohan-varma committed 2 years ago

8273912a

[TensorExpr] Add lowerings for scalar binary ops (+,-,*,/,&,|,^,<<,>>,cmp). (#71298)

Mikhail Zolotukhin committed 2 years ago

bd6ec4ef

[TensorExpr] Support scalar intermediate and output values. (#71186)

Mikhail Zolotukhin committed 2 years ago

1dbcde2a

Define check_sizes_nonnegative as inline (#71640)

peterbell10 committed 2 years ago

530e7f61

Fix symbolic shape function for `flatten` (silvasean's) (#71762)

makslevental committed 2 years ago

88c298c2

[acc_tracer] Add test coverage for retracing (#71752)

jfix71 committed 2 years ago

66939e3b

Separating CaptureDataFrame out of DFIterDataPipe (#71776)

VitalyFedyunin committed 2 years ago

b36b11cb

Obliviate ALL_TENSORTYPES and ALL_TENSORTYPES2. (#71153)

khushi-411 committed 2 years ago

dfcbe059

Change `test_conv_large` parameter initialization (#71521)

eqy committed 2 years ago

166d4e42

[cuDNN] Add a new optimized cuDNN RNN algorithm for small RNN hidden_size (#62143)

xwang233 committed 2 years ago

965b9f48

udpate missing ops message (#71294)

linbinyu committed 2 years ago

358b5078

Disable some forward mode AD tests (#71791)

anjali411 committed 2 years ago

24f577dc

[Quant] Enable default reference path for CopyNodeQuantizeHandler (#71168)

terrychenism committed 2 years ago

e4500306

Set correct device id on efficientzerotensors (#71611)

anjali411 committed 2 years ago

5dd6cd55

use legacy unrolled kernel for non-trivial offset calc cases (#71710)

ngimel committed 2 years ago

ce6e6812

Release GIL when assigning to real or imag components (#71747)

peterbell10 committed 2 years ago

f3ebf06e

Release GIL in Tensor indexing functions (#71728)

peterbell10 committed 2 years ago

dba42056

Allow torch.Tensor.real on real-valued tensors (#71718)

anjali411 committed 2 years ago

de8d0203

Check the availability of MAGMA / cuSOLVER when setting the Linalg backend. (#69826)

lezcano committed 2 years ago

03f1f0cf

Add hascuSOLVER flag to Context (#69825)

lezcano committed 2 years ago

332d67b0

`linalg.matrix_rank`: fix cpp interface + add more overloads (#70575)

nikitaved committed 2 years ago

12e01f78

edge_order check in torch.gradient only applies to dim argument (#67926)

Jonathan Colen committed 2 years ago

33403f48

Remove copies in jit_log.cpp (#67841)

nunoplopes committed 2 years ago

f3e81f3e

remove hasPrimaryContext workaround on ROCm (#71146)

rraminen committed 2 years ago

07ca1fc8

Newer Older