strisunshinewentingwang/pytorch

Pull Requests Commits

MHA optimizations (#93234)

milesial committed 1 year ago

6c555b29

[fx] fix type promotion in `binary_magic_impl` (#91376)

nkaretnikov committed 1 year ago

162e3ca5

[fx] throw exceptions on invalid input in `FloorDiv` (#93143)

nkaretnikov committed 1 year ago

34bcbfbd

[fx] test `FloorDiv` against Python impl (#93142)

nkaretnikov committed 1 year ago

ba614f3a

[fx] add SymPy assumptions to `FloorDiv` (#93185)

nkaretnikov committed 1 year ago

e7c63b96

Add count to FakeTensorMode.__torch_dispatch__ (#93936)

ezyang committed 1 year ago

2481fc0d

Short circuit device property access on FakeTensor (#93946)

ezyang committed 1 year ago

12f22655

[primTorch] Rewrite nan_to_num ref in terms of aten functions (#93952)

peterbell10 committed 1 year ago

77acb556

[primTorch] Rewrite is{,pos,neg}inf refs in terms of aten functions (#93951)

peterbell10 committed 1 year ago

72385bbd

[CI][BE] Move docker forlder to `.ci` (#93104)

malfet committed 1 year ago

6c4dc98b

cauchy_ few fixes (1) check gamma > 0 (2) better dtype error log (#93314)

min-jean-cho committed 1 year ago

6e1cfcdf

[Dynamo] modify IPEX backend (#92067)

jiayisunx committed 1 year ago

d7c71a95

Vectorized more stable complex division (#93277)

mfkasim1 committed 1 year ago

aaa27a6b

cumsum, cumprod, logcumsumexp: adjust grain size (#94025)

mingfeima committed 1 year ago

b41e2779

compute dynamic tensor shapes for indexing on the host (#93872)

ngimel committed 1 year ago

ca845084

[Fake Tensor] Allow fake meta by default, delete unused ctor args (#93993)

eellison committed 1 year ago

e4f11e01

[Inductor] Fix OpenMP discovery on MacOS (#93895)

malfet committed 1 year ago

be364c0c

[PTD] Land 'to_std' utility parser fix #93209 (#94023)

fduwjj committed 1 year ago

e98a9423

Fixed issue with --diff-branch arg in dynamo benchmarks (#93989)

Fabio Rocha committed 1 year ago

63115b70

[SDPA] Remove private version and only utilize public version (#94004)

drisspg committed 1 year ago

3df0e26e

[XNNPACK] disable ARM_BF16 and ARM_FP16_VECTOR (#94020)

kirklandsign committed 1 year ago

d996acfb

abstract vectorized reduction utils on CPU (#92284)

mingfeima committed 1 year ago

dd7d47c4

collect CPU info with collect_env.py for new issues reporting (#93899)

jingxu10 committed 1 year ago

79243516

[inductor] fix crash issue when input is a view tensor (#90150)

blzheng committed 1 year ago

a71395dd

[vision hash update] update the pinned vision hash (#94016)

pytorchmergebot committed 1 year ago

732a865c

[dtensor] add split_with_sizes op (#93957)

wanchaol committed 1 year ago

d05ec0ef

avoid unnecessary static_cast (#93898)

cyyever committed 1 year ago

bfe5e125

remove std::iterator (#93924)

cyyever committed 1 year ago

dbbcefcd

Revert "[Reland] Add sym_size/stride/numel/storage_offset to native_function.yaml (#91… (#92402)"

pytorchmergebot committed 1 year ago

f7bd5d0c

Refactor dynamo training backends (#93409)

jansel committed 1 year ago

60e8c766

Newer Older