strisunshinewentingwang/pytorch

Pull Requests Commits

add an env variable to disable addmm_cuda_lt kernel

zhaojuanmao committed 1 year ago

d9e27142

[xla hash update] update the pinned xla hash (#91087)

pytorchmergebot committed 1 year ago

2b52db9c

Revert "[cuBLAS] Fix default cuBLAS workspace size and parsing for multiple workspaces (#89027)"

pytorchmergebot committed 1 year ago

39d49dbe

Clang-Tidy: Improve ctors by removing unnecessary copies and initializations (#91538)

Skylion007 committed 1 year ago

77c2a8a1

[cuBLAS] Fix default cuBLAS workspace size and parsing for multiple workspaces (#89027)

eqy committed 1 year ago

b407d98d

Remove _ignored_param_names (#91530)

zhaojuanmao committed 1 year ago

f6136331

[BE] Move internal only non-globbed lists to OSS (#91513)

malfet committed 1 year ago

6cef5948

[cuDNN][cuDNN V8 API] Improve hot path heuristics performance in V8 (#90811)

eqy committed 1 year ago

73436af4

Rename `torchtriton` (#91539)

malfet committed 1 year ago

bc92444b

Bump protobuf from 3.20.1 to 3.20.2 in /.github/requirements (#91540)

dependabot[bot] committed 1 year ago

62713636

Cache conda and pip for IOS CI (#91359)

remidomingues committed 1 year ago

fdbbd20f

switch causal mask for is_causal flag (#91171)

mikekgfb committed 1 year ago

af589b3d

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

Older