strisunshinewentingwang/pytorch

Pull Requests Commits

CMake: Support dynamic codegen outputs (#68246)

peterbell10 committed 2 years ago

9a7732e8

Rationalize API exports in torch_python (#68095)

peterbell10 committed 2 years ago

cd9da326

Output UnionType str rep with () instead of [] (#69502)

tugsbayasgalan committed 2 years ago

829b49b8

Sparse CSR CUDA: Add block torch.addmv when mat is sparse (#68708)

IvanYashchuk committed 2 years ago

a8232ee1

skip ORT tensor in TensorIterator because it doesn't have storage (#68705)

souptc committed 2 years ago

6df7b751

[SR] Simplify memory re-use algorithm (#68302)

Mike Iovine committed 2 years ago

008469c5

Making cuda 11.5 workflows periodic (#69323)

atalman committed 2 years ago

c3096379

Add conda-forge dependency for cuda-11.5 (#69541)

malfet committed 2 years ago

baac51ff

Add Union type to TorchScript Language Ref (#69514)

gmagogsfm committed 2 years ago

358e9081

[JIT] optimize_for_inference on methods other than forward (#69367)

davidberard98 committed 2 years ago

c21169ea

[JIT] run frozen optimizations on methods other than forward (#68668)

davidberard98 committed 2 years ago

60ca6776

Sparse CSR: Implement unary ufuncs (with 0->0 correspondence) (#69292)

krshrimali committed 2 years ago

63470f94

Docs: Fix broken code syntax in autograd.rst (#69362)

rodrigobdz committed 2 years ago

1a202b0c

trt engine inspector demo (#66683)

Yinghai Lu committed 2 years ago

10229e15

[JIT] check stack size after calling operator (#68788)

davidberard98 committed 2 years ago

aa9fbb9a

[DataPipe] Small change to generation script and update to DataPipe .pyi file (#69392)

NivekT committed 2 years ago

bd8d4195

[DataPipe] Removing usage of unbatch_level from .batch interface and DataFrame (#69393)

NivekT committed 2 years ago

fdfdafd1

[DataPipe] Unifying API - removing nesting_level argument from FilterIterDataPipe (#69391)

NivekT committed 2 years ago

357160e6

[DataPipe] Unifying API - removing nesting_level argument from MapperIterDataPipe (#69390)

NivekT committed 2 years ago

4478b14e

[quant][refactor] Move pattern type definition to ao/quantization/utils.py (#68769)

jerryzh168 committed 2 years ago

9cb52327

[iOS] Add LibTorch nightly build (#69341)

Hanton Yang committed 2 years ago

976b0767

[PyTorch] Avoid no-op shared_ptr dtor when constructing tuple (#69337)

swolchok committed 2 years ago

3edf1b6c

GHA: Re enable mac json uploads (#69387)

janeyx99 committed 2 years ago

617a3bd9

[AutoAccept][Codemod][FBSourceClangFormatLinter] Daily `arc lint --take CLANGFORMAT`

CodemodService FBSourceClangFormatLinterBot committed 2 years ago

945d2e38

Set non-default backend names to lower case (#69400)

bryanmr committed 2 years ago

4670f0f2

FX: ensure node stack trace survives copying (#69368)

vkuzo committed 2 years ago

2dd46d3a

[quant][graphmode][fx] Add default_replay_qconfig for ops like reshape (#69249)

jerryzh168 committed 2 years ago

ca945d98

[JIT] Separate GPU implementation of frozen_conv_add_relu_fusion.cpp (#68149)

davidberard98 committed 2 years ago

8b1e4963

Enable build-split for all CUDA-11.x version (#69494)

malfet committed 2 years ago

e55b9397

[wip][quant][graphmode] produce reference pattern for binary ops and then rewrite to quantized op (#68229)

jerryzh168 committed 2 years ago

bd8a4a93

Newer Older