strisunshinewentingwang/pytorch

Pull Requests Commits

[Dynamo] enumerate imported names for eval_frame.py (#120778)

youkaichao committed 205 days ago

2c0c70f7

[pytorch] Support output types that are non tensors (#120804)

shruthign committed 205 days ago

ef9e8998

[inductor] Apply fx passes recursively to nested subgraphs (#120665)

aakhundov committed 205 days ago

0dbef161

Revert "[dynamo] Function => FunctionCtx for placeholder obj (#120577)"

pytorchmergebot committed 205 days ago

db1cc781

Reduce create_env log level to DEBUG (#120772)

ezyang committed 205 days ago

b2e4b621

get CommsDebugMode to work with DTensor (#118769)

bdhirsh committed 205 days ago

9e0631cc

[C10D] Add ProcessGroup op_id to track ops inside coalescing region (#120745)

wconstab committed 205 days ago

381a7ad3

[C10D] Fix pointToPoint op Flight Recording (#120270)

wconstab committed 205 days ago

f85d3a02

[C10D] Add record_id to flight recorder (#120724)

wconstab committed 205 days ago

7f4d6738

skip three pyhpc models with dynamic shape test (#120599)

leslie-fang-intel committed 205 days ago

950b4843

[DDP][PT2D] Ignore gradient sync if the gradient is not defined (#120419)

fegin committed 205 days ago

31791076

Allow str inputs in non-strict tracing (#120536)

pianpwk committed 205 days ago

ab383548

Fix guard for SUPPORTED_NODES (#120798)

aorenste committed 205 days ago

1b8bb027

[Memory Snapshot] Stop clearing history when changing context (#120436)

aaronenyeshi committed 205 days ago

aa368216

Revert "Avoid COW materialization in `at::parallel_for/parallel_reduce` (#120455)"

pytorchmergebot committed 205 days ago

86ff31c4

Revert "Add test to check that COW inputs are not materialized (#119507)"

pytorchmergebot committed 205 days ago

dbe0967a

[cuDNN] bump cuDNN-frontend submodule to 1.1.2 (#120761)

eqy committed 205 days ago

7e185277

Factor out Submod compilers (#120527)

eellison committed 205 days ago

9c9bde51

Test that tlparse understands the structured logs we output (#120658)

ezyang committed 205 days ago

5b5bcf04

Move attention kernels from meta_registrations to fake_impls (#120682)

davidberard98 committed 205 days ago

d6c20297

add a note wrt torch.nn.functional.scaled_dot_product_attention (#120668)

lancerts committed 205 days ago

50073248

Fix segfault on mac when running vulkan tests (#120337)

shubhraprakash1 committed 205 days ago

e2ee87d4

Fix `nonlinearity` arg issue in RNN (#120234)

yuanx749 committed 205 days ago

e317e39a

[FX passes] Set group/batch fusion log to DEBUG level (#120780)

yanboliang committed 205 days ago

8b22fe95

Revert "Capture non tensor arguments in record_function (#120017)"

pytorchmergebot committed 205 days ago

4903e33e

[Compiled Autograd] Introduce BackwardState capture (#120382)

jansel committed 205 days ago

01ec8df6

[C10D] Fix logic for default group=None in _set_pg_timeout (#120686)

wconstab committed 205 days ago

c016ffed

[flight recorder] record process group configuration (#120262)

shengbao-zheng committed 205 days ago

11de40f8

[inductor][Gemm] Autotune with matrix_instr_nonkdim for AMDGPU (#120742)

htyu committed 205 days ago

5aa7f864

[PyTorch Use MaybeOwned when promoting indices/offsets in embedding_bag (#120755)

swolchok committed 205 days ago

b020ee5b

Newer Older