strisunshinewentingwang/pytorch

Pull Requests Commits

[dynamo] add config for displaying all guard failures (#110927)

jon-chuang committed 1 year ago

84ad3ed7

Rever [Profiler] Improve the docstring for export_memory_timeline (#110978)

DanilBaibak committed 1 year ago

8cf1a02e

[reland] Use is_symbolic instead of testing isinstance in some place (#110676)

soulitzer committed 1 year ago

bc49b1e5

[reland] Symintify guards.cpp (#110675)

soulitzer committed 1 year ago

df9a6bca

[reland] Add symbolic singleton int (#110674)

soulitzer committed 1 year ago

3842b175

[reland] Support SingletonSymNode mul with coefficient (#110673)

soulitzer committed 1 year ago

fda0a965

Re-enable a couple of fixed tests (#110770)

eellison committed 1 year ago

fb4b9e9c

Adding Backward Support for NestedTensors and FlashAttention (#97485)

drisspg committed 1 year ago

5183760c

Updates to patch version release plans (#110952)

atalman committed 1 year ago

77e5f5d8

[Profiler] Improve the docstring for export_memory_timeline (#110949)

aaronenyeshi committed 1 year ago

52b14709

cmake: allow to build pytorch as a CMake subproject (#110373)

anthonyalayo committed 1 year ago

31611b40

[collective] Add a torch.compile + functional_collectives test (#110688)

fegin committed 1 year ago

57f6368b

Re-enable test_copy_transpose_math_view, neg_view/dce fix (#110651)

eellison committed 1 year ago

c5f06b97

AOTDispatch subclass (#104483)

bdhirsh committed 1 year ago

ba86dfcd

[inductor cpp] use c10::bit_cast to avoid violating strict-aliasing (#110809)

jgong5 committed 1 year ago

8bc04f46

[FSDP][optim_state_dict] Move local optimizer state to FSDP compute_device (#110929)

fegin committed 1 year ago

7b25c2b9

[Easy] Remove unused return type from utils (#110887)

voznesenskym committed 1 year ago

fb68aa0a

[ATen core IR] De-register `full_like` and `empty_like` as core (#110924)

SS-JIA committed 1 year ago

a4253075

Nvfuser cpp api deprecation attempt 2 (#110881)

jjsjann123 committed 1 year ago

37567fdf

Revise def of contiguity in bmm (#110811)

Valentine233 committed 1 year ago

8820dda9

[custom op] Use canonical API to constrain unbacked values (#108372)

tugsbayasgalan committed 1 year ago

35e48e26

Revert "[user errors] compulsory case names, allow multiple (#110878)"

pytorchmergebot committed 1 year ago

33403336

[vision hash update] update the pinned vision hash (#110915)

pytorchupdatebot committed 1 year ago

8891da40

Revert "[Inductor] Disallow OpOverloadPacket in ir.FallbackKernel (#110567)"

pytorchmergebot committed 1 year ago

19ecb5d0

[user errors] compulsory case names, allow multiple (#110878)

avikchaudhuri committed 1 year ago

2ae71c45

[sparse] Fix semi-structured sparse shape mismatch bug (#110420)

jcaip committed 1 year ago

f10aab03

Support Numpy ints in the `torch.nn.functional.interpolate` dtype check (#110778)

ezhang887 committed 1 year ago

468a73f0

Include rank of default PG in C++ log messages (#110623)

ezyang committed 1 year ago

de3ae93e

Move at::{Refcounted,}MapAllocator to c10 (#109881)

peterbell10 committed 1 year ago

0341deb1

[export] Update custom ops docs (#110492)

angelayi committed 1 year ago

3704bf4e

Newer Older