strisunshinewentingwang/pytorch

Pull Requests Commits

Fix c10d regression during cleanup. (#94988)

Rodrigo Kumpera committed 1 year ago

641cb424

Change test_torchinductor_opinfo.py to mark skips/xfails in a better way (#94813)

Fabio Rocha committed 1 year ago

b652577d

Upload coredump from ROCm and print the stacktrace (#94938)

huydhn committed 1 year ago

981511d0

Don't use PrimTorch decomposition for empty (#94512)

ezyang committed 1 year ago

ef5de0a4

Introduce branchless implementations of TensorImpl bools (#94473)

ezyang committed 1 year ago

2f32fd77

[PTD] Introduce tracing friendly collectives. (#93990)

Rodrigo Kumpera committed 1 year ago

e22d7912

Test inductor with stock g++ (#90710)

malfet committed 1 year ago

d0fbed76

Assume sympy is always installed (#94903)

ezyang committed 1 year ago

89e16c4f

Inductor cache clear (#94918)

voznesenskym committed 1 year ago

23b1af03

avoid extra copies in batchnorm inference by introducing a new op, _native_batch_norm_legit_no_training (#94946)

bdhirsh committed 1 year ago

68600fc7

[Dynamo] Raise warning if user has hooks installed on the module (#94848)

yanboliang committed 1 year ago

2ef66591

[inductor] Get compiler from environment variable if exists (#94926)

tbennun committed 1 year ago

bfec4965

[ONNX] Support aten::bit_wise_not in fx-onnx exporter (#94919)

titaiwangms committed 1 year ago

28e69954

[complex] nansum & nanmean (#93199)

khushi-411 committed 1 year ago

a0389681

Inductor: fix incorrect result of inplace unsqueeze (#94797)

blzheng committed 1 year ago

6ae06e49

Revert "Re-enable a FX-to-ONNX kwargs Test (#94763)"

pytorchmergebot committed 1 year ago

aa9e481e

Revert "Change test_torchinductor_opinfo.py to mark skips/xfails in a better way (#94813)"

pytorchmergebot committed 1 year ago

a049bbb1

[vision hash update] update the pinned vision hash (#94866)

pytorchmergebot committed 1 year ago

e7515538

Enable half type support for unique cpu (#91666)

zyan0 committed 1 year ago

753c33bf

Re-enable a FX-to-ONNX kwargs Test (#94763)

wschin committed 1 year ago

04b4704a

[dynamo] keep submodule's name for nn.Sequential when unroolling (#94913)

ydwu4 committed 1 year ago

4b2d1bec

[inductor] enable `test_lowmem_dropout1_dynamic_shapes` (#94884)

nkaretnikov committed 1 year ago

8c44ae2f

fix: make sure `sorter` indices are inbound in `searchsorted` (#94863)

ganler committed 1 year ago

5e1de315

Hide failing merge rule's name in the internal debugging section (#94932)

ZainRizvi committed 1 year ago

a863d5e3

[dynamo] support custom __getattr__ on torch.nn.Modules (#94658)

davidberard98 committed 1 year ago

a4085ab8

Change test_torchinductor_opinfo.py to mark skips/xfails in a better way (#94813)

Fabio Rocha committed 1 year ago

bfc0d5e2

[ONNX] Enable skipped gpt2 test (#94930)

BowenBao committed 1 year ago

3d40a86a

[BE][1/N] Add deprecate msg to Sharded Partial and Replicate Tensor (#94928)

fduwjj committed 1 year ago

b4c81867

[SDPA] Update dispatch logic to check for sm86 and head_size == 128 for flash attention (#94921)

drisspg committed 1 year ago

07bc6b95

[executorch] Add RuntimeContext to generated C++ API Signature (#94570)

larryliu0820 committed 1 year ago

41865bd8

Newer Older