strisunshinewentingwang/pytorch

Pull Requests Commits

reformat caffe2/torch/ao/quantization/quantize.py (#103823)

strisunshinewentingwang committed 2 years ago

6565f3dc

wentingwang committed 2 years ago

8d46edd6

indutor: fix issue of compute index_expr range (#103147)

XiaobingSuper committed 2 years ago

79fe3aef

inductor: fix CppTile2D bf16 store complier error for cpp backend (#103659)

XiaobingSuper committed 2 years ago

01abccf6

[CI] Fix a bug that bfloat16 is also used for dashboard training run (#103816)

desertfire committed 2 years ago

adeb63de

[Oncall][MTPG] Fix flaky test multi_threaded - test_broadcast_object_list (#103568)

wz337 committed 2 years ago

15eed5b7

[Reland][ET] Select used et_kernel_metadata only (#103705)

kirklandsign committed 2 years ago

59a01c49

add override to Caffe2 (#103795)

cyyever committed 2 years ago

f5f020ad

[Doc] Fix torch.UntypedStorage.mps() doc (#103797)

LamForest committed 2 years ago

0a7351e9

Add A Pass to Fold Tensors With a Uniform Value, match sdpa on a few models (#103600)

eellison committed 2 years ago

1b16ac74

[Quant][PT2E]Enable x86 inductor quantizer (#98730)

leslie-fang-intel committed 2 years ago

dbc8eb2a

s390x simd: ensure that vectorized complex constructor behaves same to x86 (#103426)

AlekseiNikiforovIBM committed 2 years ago

2357498a

[CI] Switch inference accuracy and performance tests to bfloat16 (#103535)

desertfire committed 2 years ago

a2988c9e

Use the new analytics ID (#103766)

albanD committed 2 years ago

918fe519

[dynamo] support FakeTensor for SYM_INT/SYM_INT_LIST/INT_LIST param in python-to-cpp argument parsing (#103448)

weifengpy committed 2 years ago

9541053c

Revert "Use hipsolver for default svd case on ROCm (#103540)"

pytorchmergebot committed 2 years ago

b34ac35b

[RPC] Check stack for emptiness in interpreter (#103327)

apach301 committed 2 years ago

750cbb29

[BE] Nested namespace in `ATen/native` headers (#103753)

malfet committed 2 years ago

f1b367c4

Better function annotations for `nn.functional` (#102918)

lkct committed 2 years ago

fd4beb7a

update multipy pin to not use install options (#103758)

PaliC committed 2 years ago

36ff9879

Do not materialize entire randperm in RandomSampler (#103339)

aviverma01 committed 2 years ago

d80174e2

Enhance decorator _use_grad_for_differentiable (#103567)

wenh06 committed 2 years ago

67babf7a

[Inductor][FX passes] Forward fix an internal unit test failure. (#103739)

yanboliang committed 2 years ago

5875a2fb

Add activation functions (ReLU and SiLU for now) for structured sparse linear operator (#101339)

alexsamardzic committed 2 years ago

8fc687f7

[gloo] Make it possible for gloo TCPStore to take over an existing socket fd (#103478)

jmswen committed 2 years ago

0da38409

[quant][pt2] Handle literal conv args in convert QAT (#103731)

andrewor14 committed 2 years ago

2bc56bec

[operator_compile_check] Add FakeTensor testing (#103595)

zou3519 committed 2 years ago

08a05464

use mutable_data_ptr for grad_input in backward passes (#98999)

ezyang committed 2 years ago

23c14340

Use hipsolver for default svd case on ROCm (#103540)

alugorey committed 2 years ago

0a4a7d4b

[RFC]: Create aten native op for constrain_range (#103346)

XuanQi committed 2 years ago

b27c3558

Older