strisunshinewentingwang/pytorch

Pull Requests Commits

Update test_cuda.py

eqy committed 1 year ago

Verified 8eb8e35d

Update aten/src/ATen/cuda/CublasHandlePool.cpp

eqy committed 1 year ago

c6f23b76

Update aten/src/ATen/cuda/CublasHandlePool.cpp

eqy committed 1 year ago

97d2b654

reduce max mem usage during test and clear

eqy committed 1 year ago

b0955315

eqy committed 1 year ago

887daac7

eqy committed 1 year ago

c703f9f3

eqy committed 1 year ago

258713b3

only bump for Hopper

eqy committed 1 year ago

e064c2ed

eqy committed 1 year ago

225da80c

Some CMake and CUDA cleanup given recent update to C++17 (#90599)

cyyever committed 1 year ago

9710ac65

Fix NumPy broadcasting in lstsq_backward (#91460)

lezcano committed 1 year ago

d5163f52

Fix NumPy-compat broadcasting in the derivative of linalg.solve (#91456)

lezcano committed 1 year ago

051d16a2

Implement PReLU in a compositional way (#91238)

lezcano committed 1 year ago

484dd400

[FSDP][optim_state_dict][8/N] Enable fully_shard optim state_dict save and load (#91234)

fegin committed 1 year ago

0e8565d1

Properly resolve source_ref when constructing shape guards (#91058)

ezyang committed 1 year ago

f8740db4

Store source, not sname, in Symbol (#91057)

ezyang committed 1 year ago

bcf15cd9

[Profiler] Fix SOFT_ASSERT test to not raise on debug builds (#91464)

aaronenyeshi committed 1 year ago

2edf589e

Drop compute capability < 5.0 in CUDA 12 (#91213)

eqy committed 1 year ago

946e5770

[torch.func] Add docs (#91319)

zou3519 committed 1 year ago

31e66ca4

(non-batch) BSR/BSC to COO performance improvement. (#91389)

nikitaved committed 1 year ago

6f034dc0

Clang-Tidy: Prevent implicit promotion in math functions (#91450)

Skylion007 committed 1 year ago

b1bdec83

Chore: fix clang warning - mismatched tags (#91455)

Skylion007 committed 1 year ago

1c3bb2fd

Perf: Apply more clang-tidy fixups to torch headers (#91445)

Skylion007 committed 1 year ago

a34a9c34

Clang-Tidy: use modern for each loops and transparent functors (#91449)

Skylion007 committed 1 year ago

553b5928

Add an option to skip loading of debug traces (#91430)

qihqi committed 1 year ago

b8ba4802

Automated submodule update: FBGEMM (#90489)

facebook-github-bot committed 1 year ago

6ec3d65b

Add out of bounds checks inside irparser.cpp and unpickler.cpp (#91401)

m4drat committed 1 year ago

3ac61065

Set a timeout value when testing multiprocess DataLoader (#91476)

huydhn committed 1 year ago

0417da22

Fix sharded_tensor test_sharded_tensor_to_cpu (#91453)

H-Huang committed 1 year ago

bc764f45

add channels last with mixed data type support for GroupNorm backward (#89485)

CaoE committed 1 year ago

5030929c

Older