strisunshinewentingwang/pytorch

Pull Requests Commits

Copy: Fix conj bit being ignored on type mismatch (#68963)

peterbell10 committed 2 years ago

917d56a7

Support Sparse CSR transpose. Fix clang-tidy warnings. (#70582)

pearu committed 2 years ago

cfc55196

Integrate multi_tensor zero_grad into Optimizer base class (#69936)

mikaylagawarecki committed 2 years ago

3a21f38a

Re-enable debug util (for LTC_SAVE_TENSORS_FILE support)

wconstab committed 2 years ago

938bf638

print error when erroring

wconstab committed 2 years ago

66136372

Add triu as codegenned lazy op

wconstab committed 2 years ago

0fd250f7

Disable normal_ since it currently diverges from eager making benchmarks report incorrect

wconstab committed 2 years ago

c2451212

Add sigmoid_backward, binary_cross_entropy* ops used by dcgan

wconstab committed 2 years ago

5a6c70f3

Disable nvfuser horizontal fusion

wconstab committed 2 years ago

93d65385

Update path to lazy_tensor lib (for CI using py37)

wconstab committed 2 years ago

e0de6b40

Run .github/scripts/generate_ci_workflows.py

wconstab committed 2 years ago

b53b5be2

Fix missing aten symbols after merging master

wconstab committed 2 years ago

c887d393

Fix docstring for nn.Tanh (#70577)

lithuak committed 2 years ago

7627d166

Copy: Avoid extra dispatch in type-mismatch case (#68950)

peterbell10 committed 2 years ago

a5726fa9

Bazel: Only run ATen codegen once (#70147)

peterbell10 committed 2 years ago

6bc1e9d4

DOC: Merge extraheader block from theme instead of override (#70187)

mattip committed 2 years ago

daf8f79f

dbr quant: stop calling eager quant convert (#70247)

vkuzo committed 2 years ago

b331660c

dbr quant: break up test class into multiple classes (#70246)

vkuzo committed 2 years ago

479cbecc

dbr quant: start recording ops which are not quantizeable (#70200)

vkuzo committed 2 years ago

e1f2065d

dbr quant: start reusing static quant module mappings (#70196)

vkuzo committed 2 years ago

8c7ce1a6

Fix generating files for Vulkan on Windows (#69696)

mattjgalloway committed 2 years ago

7448610b

[caffe2] don't use __FUNCSIG__ when building for Windows with clang (#70561)

h-friederich committed 2 years ago

28eb3662

[PyTorch] Optimize no input NVTX collection (#70133)

chaekit committed 2 years ago

64e30f6a

NNAPI: Add qint16 support via int16 (#70621)

Akshit Khurana committed 2 years ago

e8ce01cc

[PyTorch][Static Runtime] Fast aliasing in select_tensor by manual borrowing (#68122)

swolchok committed 2 years ago

eb9950aa

[PyTorch][Static Runtime] Support memory planning for torch.to() w/o requiring copying (#67223)

swolchok committed 2 years ago

a32450bd

[PyTorch][Static Runtime] Avoid managed output tensor DCHECK (#67221)

swolchok committed 2 years ago

5407963f

[PyTorch][Static Runtime] Fix dtype changing between iterations for to() (#67394)

swolchok committed 2 years ago

5b85f0dc

Support cloning CSR tensors (#70581)

pearu committed 2 years ago

006780b0

Change to use current LLLVM APIs (#70625)

modiking committed 2 years ago

8b5f1fe7

Newer Older