strisunshinewentingwang/pytorch

Pull Requests Commits

Remove native_functions.yaml dependency from some reduction operators (#64173)

peterbell10 committed 2 years ago

9984f4bb

fix normal with empty std (#66463)

ngimel committed 2 years ago

ee38a467

Use shared CUPTI by default (#65401)

ezyang committed 2 years ago

8b0eae5a

Back out "Revert D30710710: [Pytorch Edge] Support profiling kineto events from external source" (#66421)

kimishpatel committed 2 years ago

c6216b2a

[jit] Eliminate malloc & recursive refcount bumps in HashType::operator() (#65348)

swolchok committed 2 years ago

d7916e37

[jit] Compare object identity first in ClassType::operator== (#65347)

swolchok committed 2 years ago

47c531b6

remove is_reference from all is_output_quantized (#66456)

rahxephon89 committed 2 years ago

17e79bc7

[fx2trt] open source tests for acc tracer (#66302)

jerryzh168 committed 2 years ago

702fb1de

Upgrade onnx submodule to 85546f8c44e627f8ff1181725d03cc49f675e44f (#66427)

houseroad committed 2 years ago

a6eec0c6

[FX] fuse permute021 linear pass for trt lowering (#66362)

Yinghai Lu committed 2 years ago

e6261083

Fix lstsq to work with inputs that require grad (#66426)

IvanYashchuk committed 2 years ago

8818dda2

Remove native_functions.yaml dependency from PointwiseOps (#64172)

peterbell10 committed 2 years ago

213ac4e5

Remove native_functions.yaml dependency from PowKernel (#64171)

peterbell10 committed 2 years ago

8674a3c6

Remove native_functions.yaml dependency from unary ops (#64170)

peterbell10 committed 2 years ago

1841f76c

[DataPipe] Fix HttpReader IterDataPipe Issue with streaming (#66432)

NivekT committed 2 years ago

71e17d98

[TensorExpr] Fix lowering for aten::t. (#65859)

Mikhail Zolotukhin committed 2 years ago

5f151860

[TensorExpr] Fix lowerings for aten::view and aten::reshape. (#65852)

Mikhail Zolotukhin committed 2 years ago

6864146f

[TensorExpr] Use schema instead of op name in NNC lowerings. (#65843)

Mikhail Zolotukhin committed 2 years ago

60a2a295

[TensorExpr] Nuke TE shape inference. (#65554)

Mikhail Zolotukhin committed 2 years ago

24b9b304

[Pytorch Edge] Improve bundled inputs name error handling (#65856)

JacobSzwejbka committed 2 years ago

18e46881

Revert D31386275: Migrate THCState to ATen

suo committed 2 years ago

2d155282

[PyTorch] Split c10 Type.cpp into two files to allow targets to include one of them (#66445)

larryliu0820 committed 2 years ago

d8532e35

[deploy] fix oss build (#66347)

suo committed 2 years ago

07ec250f

Fix batch_isend_irecv tests for err case (#63112)

zasdfgbnm committed 2 years ago

9a85167d

[FX] Fix issue where GraphModule.delete_all_unused_submodules deletes submodules from called leaf modules (#66430)

James Reed committed 2 years ago

3eb94436

Migrate THCState to ATen (#65948)

peterbell10 committed 2 years ago

a6774d6e

Call `PyArray_Check` only if NumPy is available (#66433)

malfet committed 2 years ago

e7b5712c

Quantization docs: add pages for Numeric Suite (Eager and FX) (#66380)

vkuzo committed 2 years ago

565cf47a

Improve quantization API docs (#66379)

vkuzo committed 2 years ago

8b125869

Fix type checking errors in torch/quantization/fx/qconfig_utils.py (#66428)

EshikaShah committed 2 years ago

88ed93c2

Newer Older