strisunshinewentingwang/pytorch

Pull Requests Commits

[aotinductor] Relax ExternKernel kwargs checking (#111167)

desertfire committed 1 year ago

73d288fd

[FSDP] fix: fix for fsdp zero2 validation error (#110139)

Edwiv committed 1 year ago

5caf2e55

[BE] Compile FBGEMM with ASAN (#111266)

malfet committed 1 year ago

6dc54fe8

update the dispatch of some operators which accept scalar (#110918)

alanhe151220037 committed 1 year ago

cff8bf47

[TP] Add prepareInput and output for input/output DTensor layout annotation in the parent module in TP API (#111166)

fduwjj committed 1 year ago

8085e08a

[state_dict][3/N] Cleanup StateDictOptions, make it more readable (#111275)

fegin committed 1 year ago

7c67139e

[TP] Refactor Parallel Style to make it more usable (#111160)

fduwjj committed 1 year ago

3a8b10e2

Intra-graph reordering pass on Inductor scheduler IR (based on #100762) (#108091)

yf225 committed 1 year ago

b28cb43f

[2/N] Apply clang-tidy to aten/src/ATen/core/ (#111006)

cyyever committed 1 year ago

8bd5eb8c

Fix size_hint call sites failing on unbacked SymInts (#110520)

aakhundov committed 1 year ago

d7317d8a

[inductor] Allow backend compiler to skip (#111153)

bertmaher committed 1 year ago

0013611c

[BE] Move ASAN from clang-12 to clang-15 (#111218)

malfet committed 1 year ago

48e4d183

Revert "[state_dict][3/N] Cleanup StateDictOptions, make it more readable (#111108)"

pytorchmergebot committed 1 year ago

581d97c1

[export] Use meta val from the old nodes in run_decompositions(). (#111225)

zhxchen17 committed 1 year ago

11ac4ace

[sparse] semi-structured sparse + torch.compile support (#111049)

jcaip committed 1 year ago

ac02531b

Add `lazy_clone_storage` to create COW storages (#110192)

kurtamohler committed 1 year ago

1c308144

Revert "Add `lazy_clone_storage` to create COW storages (#110192)"

pytorchmergebot committed 1 year ago

48278240

Add constant node sizes to proto size calculation (#111097)

Thiago Crepaldi committed 1 year ago

4f4e2c1c

[dynamo] collapse local and global guard builders (#111226)

voznesenskym committed 1 year ago

3b08a4a6

Skipped CUDA Flags if C++ Extension Name includes "arch" Substring (#111211)

gauenk committed 1 year ago

bb89a9e4

[inductor] Adding a way to force fusion of int_mm with mul (#111125)

HDCharles committed 1 year ago

f4297576

Documentation Clarification on torch.compile Example (#110942)

yewentao256 committed 1 year ago

c1511633

[BE] Enable Ruff's Flake8 PYI045 (#111184)

isdanni committed 1 year ago

00d96263

[export] Update serialization schema to input/output specs. (#845) (#111204)

zhxchen17 committed 1 year ago

ba7b9211

Fix torch.diagonal for torch.onnx.export when dim1<0 or dim2<0 (#111130)

veritas-Qiu committed 1 year ago

a3e9b800

Un-skip a bunch of UnaryUfuncInfo bfloat16 tests (#110799)

int3 committed 1 year ago

375e7bd0

Update arg{min,max} tests and docs (#110845)

int3 committed 1 year ago

d8de45d2

Don't sympify reflection_pad2d ranges (#111212)

ezyang committed 1 year ago

d38472c1

[BE] Enable Ruff's Flake8 PYI034 (#111105)

isdanni committed 1 year ago

382327bd

Allow strided layout in torch.normal (#111205)

lezcano committed 1 year ago

2fd546aa

Newer Older