strisunshinewentingwang/pytorch

Pull Requests Commits

Add dimension check in tensordot (#94497)

yanbing-j committed 2 years ago

d9f822b5

[optim] Widen the cases for defaulting to foreach (#95820)

janeyx99 committed 2 years ago

75cb99e5

[optim] include nn.Parameter as foreach supported (#95811)

janeyx99 committed 2 years ago

2bcf863f

[vision hash update] update the pinned vision hash (#95843)

pytorchmergebot committed 2 years ago

45fd1b39

[sdpa] Add broadcasting for batch and num_heads dimensions to fused kernel nested preproc (#95657)

mikaylagawarecki committed 2 years ago

4973ca5e

[inductor] Improve error messages (#95567)

jansel committed 2 years ago

62b77558

Bump version 2.0.0 -> 2.1.0 (#95790)

atalman committed 2 years ago

d1ec9a51

[MPS] Remove casts from reduction/cumsum/sort ops starting with macOS 13.3 (#95817)

DenisVieriu97 committed 2 years ago

4d3352ed

Small doc update for torch_compile_debug (#95809)

mlazos committed 2 years ago

184fb9f1

[3/3] Update `.pyi` Python stub files and enable `'UFMT'` linter (#95268)

XuehaiPan committed 2 years ago

1fd11994

Fix typos in documents under torch directory (#95709)

kiszk committed 2 years ago

b3d8fae0

[JIT] Improve source attribution for NamedTuple type inference (#95761)

davidberard98 committed 2 years ago

b35e6714

[dynamo] Fix OrderedDict reconstruction bytecode (#95800)

williamwen42 committed 2 years ago

053205aa

fix some tiny code issues (#95757)

cyyever committed 2 years ago

6786a24f

Remove mention of dynamo.optimize() in docs (#95802)

msaroufim committed 2 years ago

f7b26bdd

Don't use guardless contiguity/stride-like implementations (#95733)

ezyang committed 2 years ago

deaf077d

Apply peephole for eval mode when constant folding is enabled only (#95801)

Thiago Crepaldi committed 2 years ago

a9a3a1bd

Always get attr static out (#95771)

voznesenskym committed 2 years ago

8093abce

Rename func (#95639)

voznesenskym committed 2 years ago

34a7c79e

[ROCM] Restrict pytorch rocm to only use triton 2.0.x (#95793)

dllehr-amd committed 2 years ago

de86538f

Revert "Enable thp(transparent huge pages) for buffer sizes >=2MB (#93888)"

pytorchmergebot committed 2 years ago

2936c8b9

Update inductor-perf-test-nightly.yml (#95807)

weiwangmeta committed 2 years ago

13340638

[sdpa] move seq_len_1 check and replace with seq_len_0 check in sdp_utils (#95486)

mikaylagawarecki committed 2 years ago

63796d35

Clean up install_triton and install_filelock in CI (#95754)

huydhn committed 2 years ago

72b9d45e

Preserve specialize_int_float during export (#95741)

tugsbayasgalan committed 2 years ago

dd889545

[MPS] Add fixes for div with floor and raise error for div_trunc (#95769)

kulinseth committed 2 years ago

5d9d8c61

Retry Merge: extract utils from check labels ptr (#94899)

Goldspear committed 2 years ago

5ba4dafc

Logaddexp for complex in CPU (#95717)

mfkasim1 committed 2 years ago

975333d8

[EASY] Make has_hint work on more things than just SymInt. (#95792)

ezyang committed 2 years ago

97fbceea

[CI] Increate the timeout limit for benchmark test (#95787)

desertfire committed 2 years ago

879f0c3f

Newer Older