strisunshinewentingwang/pytorch

Pull Requests Commits

[executorch] Run llama in xplat (#118831)

lucylq committed 233 days ago

6c8faf46

Move lr tensor to cuda if needed (#119073)

mlazos committed 233 days ago

a64b03a5

[dynamo] Fix incorrect docstring placements in _guards.py. (#119114)

ttsugriy committed 234 days ago

41b63b26

Remove extra graph breaks (#118987)

mlazos committed 234 days ago

9a8e3b07

[FSDP] Fixed `device_mesh` and auto wrap (#119064)

awgu committed 234 days ago

ce40ee8e

[MPS][BE] Add native lerp support (#119036)

malfet committed 234 days ago

18fc1ca7

Inline gradcheck functions since they don't have C bindings (#119047)

jamesjwu committed 234 days ago

30d3ff16

ProcessGroupGloo::reduce_scatter_tensor_coalesced (#118911)

yifuwang committed 234 days ago

372e9550

[c10d] add an unit test for unordered destruction of PGs (#119045)

shuqiangzhang committed 234 days ago

65314a61

Change the internal assert to torch_check in torch::nn::functional::InterpolateFuncOptions (#117831)

lancerts committed 234 days ago

857508fa

Document file format returned by torch.save (#118719)

mikaylagawarecki committed 234 days ago

9ffed223

[dynamo] decrease logging level for graph break in higher order op. (#119079)

ydwu4 committed 234 days ago

2eba82d1

[submodule kineto] Enable profiler connection to daemon during init for cpu only jobs (#118320)

briancoutinho committed 234 days ago

d91d21fd

[DCP][BE] Let FsspecWriter and FsspecReader inherit from FileSystemWriter and FileSystemReader (#118887)

fegin committed 234 days ago

494c2ec0

Enable scaled_mm on sm89 devices (#118881)

drisspg committed 234 days ago

6b009ace

[auto_functionalize] Remove mutated_args_name from args (#119050)

angelayi committed 234 days ago

440b7d52

Revert "Remove parent device mesh check (#118620)"

pytorchmergebot committed 234 days ago

3aeaa21e

Expose aggressive_recomputation as an inductor config (#118943)

Yuzhen11 committed 234 days ago

de6a9060

Improve example for ``torch.mode()`` (#115308)

linus-md committed 234 days ago

7bbd9bef

[inductor] make multi-kernel work with cpp-wrapper (#117813)

shunting314 committed 234 days ago

c24ffc3f

Add torch check for dtype within bilinear (#118900)

lancerts committed 234 days ago

576383c2

Revert "Add --filter-rank to torchrun: allow logs filtering by rank (#118562)"

pytorchmergebot committed 234 days ago

a4355d6b

[c10d] logging utility for cpp-python stacktrace (#118924)

weifengpy committed 234 days ago

63fd6883

[ONNX] Eliminate redundant TODOs (#119060)

titaiwangms committed 234 days ago

a3cec6a7

[export] Prevent specialization on backends (#118683)

angelayi committed 234 days ago

454e6b38

[export] fix forward test_lift_unlift (#119090)

suo committed 234 days ago

db2225da

[dynamo] bypass graph break due to masking if inference mode (#119056)

ydwu4 committed 234 days ago

9fe3693b

[fx] fix for subgraph rewriter (#119052)

suo committed 234 days ago

4d45c68c

[DeviceMesh] Alllow 1d slice from 1d mesh (#118895)

wz337 committed 234 days ago

c908caf9

[dynamo][higher order ops] Remove restore side effects logic (#118420)

anijain2305 committed 234 days ago

6379010e

Newer Older