strisunshinewentingwang/pytorch

Pull Requests Commits

[Reland] Avoid saving self for `softmax` and `log_softmax` (#66018)

soulitzer committed 2 years ago

4cdfcedd

Refactor functional api vectorized jacobian to use batched grad parameter (#65566)

soulitzer committed 2 years ago

8f5631b8

Add batched_grad parameter to `autograd.grad` (#65564)

soulitzer committed 2 years ago

73901b09

Allow None to pass through for vmap (#65565)

soulitzer committed 2 years ago

b6d5f1ee

[Static Runtime] Fix bug of creating output aliases in aten::embedding_bag (#65516)

d1jang committed 2 years ago

89ed9bda

Fix LLVM-12 UB in generate_proposals_op.cc (#66009)

r-barnes committed 2 years ago

40948a93

Added validation of mode parameter in AveragedModel (#65921)

prabhat00155 committed 2 years ago

c7748fc1

[gpu ne eval] disable adam decay unit test for gpu (#66056)

Hector Yuen committed 2 years ago

0fc6bd2e

Back out "[caffe2] fix LLVM-12 nullptr-with-nonzero-offset UBSAN error" (#66055)

mehtanirav committed 2 years ago

29c0725e

[WIP] skip constant folding dequant node (#63991)

Krovatkin committed 2 years ago

7c529633

selective trt import based whether we have gpu or not (#66045)

Yinghai Lu committed 2 years ago

8a307640

[PyTorch Edge] Skip writing version during backport (#65842)

cccclai committed 2 years ago

8b8012a1

[JIT] Selectively enable precise alias analysis for TupleConstruct (#66025)

d1jang committed 2 years ago

7941590a

Revert D31326599: [pytorch][PR] Compile without -Wno-unused-variable

malfet committed 2 years ago

e4ee5ca6

Revert D31359010: [pytorch][PR] Fix cang-tidy regressions caused by #65954

malfet committed 2 years ago

5ef350d7

Fix cang-tidy regressions caused by #65954 (#66040)

malfet committed 2 years ago

c269f471

Fix nll_backward for negative weights (#64572)

lezcano committed 2 years ago

ca76e193

[XROS][ML] System specific adjustments for UTs to work. (#65245)

Karol Kosik committed 2 years ago

eb3b9fe7

GELU acc OP (#65957)

FusionPower committed 2 years ago

363ccb25

Compile without -Wno-unused-variable (#65954)

malfet committed 2 years ago

a6280ab6

Fix shape inference dim_type for Clip, Mean, Div (#65996)

qxy11 committed 2 years ago

10f62942

model_dump: Fix memory computation when both constants and data tensors are present (#66006)

dreiss committed 2 years ago

e1d963e8

model_dump: Add a helper to produce html with a single call (#66005)

dreiss committed 2 years ago

23caeb3f

Upload test failures to RDS (#65873)

driazati committed 2 years ago

d9a95e66

[fx2trt]add support for torch.tile (#66016)

842974287 committed 2 years ago

f85d7422

Forward fix type hint for DataLoader (#66001)

ejguan committed 2 years ago

060e41ea

Revert D30634700: [pytorch][PR] Fix typo in tensor docs

suo committed 2 years ago

ad889d0b

[fx-acc] add acc_op optimization flags and decorator (#65928)

alexbeloi committed 2 years ago

7d220079

Fix typo in tensor docs (#64160)

Akshit Khurana committed 2 years ago

d9374737

Re-generate workflows (#66027)

driazati committed 2 years ago

8e869528

Newer Older