strisunshinewentingwang/pytorch

Pull Requests Commits

[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

[PyTorch] fix mixed int32/int64 indices/offsets for embedding_bag_out (#120752)

swolchok committed 205 days ago

98d15294

[codemod][lowrisk] Fix deprecated use of 0/NULL (#120740)

emmettneyman committed 205 days ago

db925582

Let torch dynamo inline torch.func.grad (#118407)

guilhermeleobas committed 205 days ago

491c2b46

derived dim (#118729)

avikchaudhuri committed 205 days ago

54729239

TransformerEncoder/Decoder: add type hints (#120550)

adamjstewart committed 205 days ago

9c55aa6f

Update flash_attention kernel from 2.3.6 to 2.5.5 (#118935)

drisspg committed 205 days ago

4b7a5218

Revert "Increased compile time max GPUs to 512. Switched to int16_t DeviceIndex. (#119639)"

pytorchmergebot committed 205 days ago

a9d9077f

Newer Older