strisunshinewentingwang/pytorch

Pull Requests Commits

Make LazyGraphExecutor extensible (#87218)

antoniojkim committed 1 year ago

d37dc6f6

Fix typo under torch directory (#87274)

kiszk committed 1 year ago

d80a5f9a

[MPS] Revamp copy_to_mps_ implementation (#86956)

malfet committed 1 year ago

ae62cf7c

[dynamo] [easy] RM spurious `)` (#87439)

voznesenskym committed 1 year ago

435e78e5

Python binding for dispatcher getAllOpNames (#87422)

SherlockNoMad committed 1 year ago

ab901b48

[inductor] Fix channels_last conv2d propagation when CuDNN is not found (#87266)

soumith committed 1 year ago

7caeac17

Fix registration hooks (#87369)

antoniojkim committed 1 year ago

6b59d9b5

[AOT][CUDAGraphs] torchdynamo -> torch._dynamo (#87243)

soumith committed 1 year ago

ff43288d

[functorch] fix AOTAutograd tutorial (#87415)

zou3519 committed 1 year ago

13ab8193

Enable inductor CI for huggingface (#86792)

desertfire committed 1 year ago

b1cf377c

[Inductor] Convert 0d CPU tensor to scalar during triton codegen (#87329)

yanboliang committed 1 year ago

9ba63225

Add `weights_only` option to `torch.load` (#86812)

malfet committed 1 year ago

961ebca2

Remove jansel/voz from dynamo CODEOWNERS (#87430)

jansel committed 1 year ago

e3d73bbb

Improve the performance of validate_non_overlapping_shards_metadata (#85639)

fegin committed 1 year ago

bd1e95ce

Back out "Revert D40198461: [pytorch][PR] Backport currently dont work with some models if:" (#87124)

qihqi committed 1 year ago

a42fbfa0

Revert "[dynamo] use optimizers correctly in benchmarking (#87311)"

pytorchmergebot committed 1 year ago

f38a88c4

[Inductor] TorchInductor tracing fx_graph.py should import overrides (#87271)

yanboliang committed 1 year ago

a91abedf

set ci in mps (#87325)

clee2000 committed 1 year ago

1801b57c

Unify decomp registries into global_decomposition_table (#86857)

SherlockNoMad committed 1 year ago

f7da9db9

Add General Project Policies (#87385)

Svetlana Karslioglu committed 1 year ago

7e83f65a

[maskedtensor] fix docs formatting (#87387)

george-qi committed 1 year ago

17202b36

add deprecation warning to nn stateless functional_call (#87367)

samdow committed 1 year ago

bc8cf332

[ci] handle libomp upgrade on github (#87382)

clee2000 committed 1 year ago

9b88dcf2

[functorch][docs] Downgrade the warning about forward-mode AD coverage (#87383)

zou3519 committed 1 year ago

08268639

[dynamo] Add support for invoking nn sequential (#87156)

voznesenskym committed 1 year ago

2fd008ed

Fixed tune_layout to not do anything for non-2d convolutions (#87328)

Chillee committed 1 year ago

68e946b0

[functorch] Fix torch.cat batching rule (#86932)

zou3519 committed 1 year ago

b805e1ab

[Profiler][Trivial] Small style and safety fixes (#86752)

Taylor Robie committed 1 year ago

c16b7b41

[dynamo] avoid popen.communicate() (#87335)

zdevito committed 1 year ago

1e4a2742

Retry sccache downloads (#87306)

ZainRizvi committed 1 year ago

75a5a46a

Newer Older